December 15, 2020 • 55 mins
In one sentence...
Human health upgrades may be unlocked by a data storage expert focused on probing the proteome.
Overview:
It’s not everyday that one hears about the crucial relationship between digital data storage and biology, yet this is core to Sujal Patel’s work leading Nautilus Biotechnology. By any measure already an accomplished tech entrepreneur, Sujal’s experience building Isilon Systems (acquired by EMC for $2.5 billion) became a perfect, if non-obvious, ingredient for disentangling the complexities of proteins.
If successful in analyzing the proteins within thousands of cells, Patel, his co-founder Parag Mallick, and their team may discover new drugs for present and emerging diseases. Pioneering the birth of this new proteome sequencer instrument involves deep and broad interdisciplinary collaboration. Patel says, “On our staff are mechanical engineers, software engineers, electrical engineers, and biophysicists working side-by-side with biochemists, organic chemists, bioengineering majors – all these disciplines come together to build a complete solution.”
Topic Timecodes:
[00:00] Episode preview
[00:59] Nautilus Biotechnology and why the need to understand what’s going on at the protein level
[05:03] Isilon to Nautilus
[11:23] Early stages of Nautilus
[14:13] Tough stuff at Nautilus
[19:13] Goal of Nautilus
[21:14] How Sujal went about learning a completely new field
[24:17] What exactly is Nautilus building
[26:54] Broader landscape of proteomic analysis companies
[32:17] What is unique about Nautilus Bio
[35:53] Nautilus Bio’s relationship with customers
[38:22] Nautilus and COVID-19 / other pandemics
[41:37] Going faster at Nautilus due to high demand
[43:51] Adjusting to life with COVID-19 at Nautilus
[48:02] Priorities for Nautilus in the near-term
[50:58] What’s been the most fun for Sujal in building Nautilus?
[53:36] Shameless plugs
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Sujal Patel (00:01):
In order to make a dent in the world that we're
going after, you have to measuremillions and billions of
molecules in a single run of aninstrument, and that can't take
more than a few days, and thereason you have to measure that
much is that your cells, eachcell has a lot of proteins in
it.
Announcer (00:19):
Welcome to Tough Tech Today with Meyen and Miller.
This is the premier showfeaturing trailblazers who are
building technologies today tosolve tomorrow's toughest
challenges.
Jonathan (00:33):
Welcome to Tough Tech Today with Meyen and Miller.
Today, we have the honor ofspeaking with Sujal Patel,
co-founder of NautilusBiotechnology. Welcome, Sujal.
Sujal Patel (00:48):
Thank you, Jonathan and Forrest. Thanks for having
me.
Jonathan (00:50):
To get our audience up to speed with what you're
working on, could you tell us abit more about Nautilus
Biotechnology?
Sujal Patel (00:59):
Sure, Nautilus Biotechnology is a incredibly
interesting problem between thebiological domain, physical
sciences domain, and the techdomain. You know, if I'm to
explain the company in storyform, what I'd say is, is that
if you look at biotechnologyover the course of the last few
(01:20):
decades, and specifically welook at genomics, we as humanity
conquered genomics over thecourse of the last two decades
today, if I take a drop of yourblood, and I want to understand
what your genome is, I can getyou 99% of your genome for
$1,000. In a couple of days,it's a complete commodity. But
the issue is, is that yourgenome really doesn't change
(01:43):
from the day that you're born,the day that you die, it
actually is very static, itdoesn't contain anything real
time information about whetheryou're sick today or what's
going on. And because of that,there is this big move to try to
understand more of what's goingon in your cells, and your cells
are all made of proteins. And ifyou really want to understand
what's going on inside of humanbody, you have to get to the
(02:04):
protein level. And the issuethat exists in the world today
is that the very best techniquesthat we have to analyze proteins
from that same drop of blood orfrom any biological sample are
very different than genomics. Inthe genome, take a drop of
blood, $1,000, in a few days, Iget 99% of the answer. If I want
(02:25):
the proteome, which is themakeup of the proteins, I could
take that same drop of blood,the very best techniques on the
earth, leveraging complex piecesof equipment, like mass
spectrometers would take $50,000in a month to analyze that
sample. And at the conclusion ofthat, we would have identified
8% single digit 8% of theproteins that are the sample.
(02:47):
And so Nautilus was founded totry to improve that
dramatically.
Forrest Meyen (02:52):
So can you walk us through like exactly why we
would want to know what's in theparticular proteins? Because I
imagine your DNA encodes for allof those proteins to allow them
to exist? Are you trying to getinformation on environmental
factors or different things fromthose tests?
Sujal Patel (03:14):
It's much more than just environmental factors. But
I think that Forrest you're onthe right track there. So
obviously, your genome encodesall of the possible proteins
that could be created, but for awide variety of reasons, what's
occurring in your cells, and theproteins in your cells do not
match in any particular way whatyour genomic if we look at it
(03:37):
in software terms, what yoursoftware code says. And that's
the reason why if you took twotwins, and you let them age to
their 40, and one took care ofthemselves and one ate potato
chips all day and neverexercised, they'd look
completely different. Theenvironmental factors, the
health factors, what's going onin that particular body, all
(03:57):
those things are encoded in yourproteins. This is the reason why
if you looked at the FDAapproved drugs, over 90% of the
FDA approved drugs targetproteins. Most of our
diagnostics target proteins. Andone of the side effects of this
inability to measure proteinseffectively is that drug
(04:19):
development has gotten lessefficient, and it's gotten
worse. So over the last twodecades, if you looked at the
number of new FDA approved drugsevery year, it's basically dead
flat. And that's in the face ofa quadrupling of the global R&D
spend in pharma from about 50billion to 200 billion so we've
(04:39):
quadrupled spend, but FDAapprovals is flat. Because of
that, only two out of 12marketed drugs that reach the
market actually return the R&Ddollars that went into the
advocacy of the timelines havebecome worse and extended. And
not understanding the proteinswhich make up all of your cells
(05:00):
is a huge part of that problemthat's that's been created here.
Jonathan (05:03):
This is a... really complicated problem set. And so
I'm curious with your backgroundwith as CEO of Isilon Systems
which from my understanding wasfar and away from a biotech
company. And so could you tellus a bit... how to connect those
(05:28):
dots between your work withIsilon and then now find
yourself at the helm of a reallyexciting biotech company.
Sujal Patel (05:36):
Yeah, it's a great question, Jonathan. Because the
path for me to get from my lastcompany to here is a really
interesting path. And it's oneof those sort of small world
types of stories. So my lastcompany... well, let me kind of
back up. Up until four yearsago, my entire career has been
in the tech world. And the lastcompany that I spoke to was
(06:00):
company called Isilon, which wasa company founded in January of
2001, went public in 2006. Andthen we eventually sold it for
2.6 billion in 2010. Thatcompany was focused on building
a new storage architecture forunstructured information, things
like digital content, video andimages machine generated data.
(06:21):
Two decades ago, storage systemsof the time, were really focused
on string text based informationand databases and credit card
transactions. And so we built ascalable architecture. And we
took that to the world focusedon different verticals that were
undergoing a digitaltransformation. So we started
selling into companies doingphoto sharing on the internet,
(06:42):
into film broadcasts andtelevision which was undergoing
a transition from analog todigital. And then we moved into
manufacturing and semiconductorand by '04 and '05, life
sciences and biomedical researchbecame a really big market for
us. And this revolution thatoccurred in genetic sequencing
was a big part of the successthat we had in that vertical
(07:03):
market. Back in 2004, which is16 years ago, I met a guy named
Parag Malik and Parag became avery large customer of mine at
Isilon supporting his proteomicsResearch Lab at Cedars Sinai
Medical Center, and Parag 16years later, is my co-founder
(07:23):
here at Nautilus. And that, ourpaths intersected many times
over the course of those 16years. First, Parag was in the
role of a customer to me, and Igot to know Prague really well
and thought he was one of thesmartest guys that I ever met.
About nine years ago, Parag lefthis role at Cedars Sinai and
(07:47):
went to Stanford to teach andstart a research lab, which was
really building technologies topursue a personalized medicine,
with a focus of sitting at theintersection of tech and biotech
products, kind of a uniqueanimal. And he has academic
degrees in both biochemistry andcomputer science. And my wife
and I were so impressed with thework that he was trying to do in
(08:10):
his lab and then he has donethat we decided to
philanthropically support hislab. And we've done that for the
last nine years, which built areally close relationship
between Parag and I, and thatled to Parag calling me up in
2016 and saying, Hey, I'm gonnastart a company because I have
an idea and let me tell youabout it. And what started as a
(08:31):
conversation, soliciting advicevery quickly turned into two
founders going off and trying tosolve a really tough, really
hard problem. And it's been anamazing ride for four years of
kind of moving from the crackpotidea phase through feasible and
now into real engineering.
Jonathan (08:50):
That's so cool to hear how it's like an organic
relationship that developed,that you found that you could...
worked well together because ofgo founding a company is
non-trivial, and absolutely astrain on a relationship. What
were you doing when you got thatcall in 2016? And were you
(09:11):
looking for your next bigcompany to start?
Sujal Patel (09:15):
Yeah, so it's interesting, right. So we sold
Isilon in December of 2010. Andwe just closed our 100 million
dollar quarter. And I hadcommitted to the acquiring
company that I would grow thebusiness to a billion dollar run
rate and then I would leave. Wewere in a very high growth
rate that only took 23 months,and then I left. And I did some
(09:36):
consulting for them to help themthrough a few other things that
were going on. But largely, Istarted to think about what I
wanted to do next. And I hungout with venture capitalists, I
made a lot of investments incompanies that total I've made
about 80 investments in the last10 years in private companies. I
had six boards at some point andI started to work with
entrepreneurs and about sixmonths in I realized that I
(10:03):
couldn't be on boards andinvest; I really was too young,
too passionate, too excited togo and actually build something.
And I knew I wanted to go outand do something again. But I
also knew that I didn't want todo enterprise software,
enterprise hardware, I didn'twant to do something that was
like what I did before. And Ialso had this burning desire to
go work on really hard problemsthat have big societal impact.
(10:26):
And so I was only really lookingin three areas I was looking in
healthcare technology, I waslooking in clean energy, and I
was looking at a couple of ideasin consumer, which would have
very broad implications. Andthat's it. And none of those
ideas ever graduated to thepoint where I said, that was the
one for me until Parag came andit literally was that first one
(10:48):
hour call with Parag, I wentthrough it, I'm like, okay, if
anyone were to bring me an idealike this, I'd call Parag and
say, Parag, what do you thinkabout this idea? And then I said
to myself, Parag is going to goput his Stanford career which is
really successful on hold, ifthis was an incredibly exciting
in the opportunity of alifetime. And I got there in one
(11:10):
hour, obviously, there are a lotmore conversations from there,
but in one hour, I was like,Okay, this is it like this is
that thing. And it's been it'sbeen super exciting, getting
into a whole new and differentspace.
Forrest Meyen (11:23):
So you called the early phase of the company, or
the crackpot idea phase. Now wasit really a crackpot idea? Did
you shift a lot into what it isnow? Or were you just getting
more information to gainconfidence that you had
something really special?
Sujal Patel (11:43):
So that's an interesting question, right? I
mean, tongue in cheek, we callthis the crackpot idea phase in
2016. But what I will tell youis that the idea was really
unformed at that point. But wehad an idea that the method that
Parag was proposing to analyzehuge numbers of proteins and get
(12:04):
to a very specificidentification with those
proteins. And we an idea thatthat was possible, but there
were a huge number of unansweredquestions. And a big chunk of it
was related to the corealgorithm that Parag conceived
of that is the underpinnings ofwhat we do. And a big part of it
is how do we scale this up anddo it reliably at massive scale
(12:28):
at low cost. And now, that part,is really the bulk of what we
took the last four yearsworking... what we've done over
the last four years. The firstsix months were algorithmically
moving it from crackpot idea to,okay, this really can happen.
And in all honesty, I said tomyself, well, let's go see what
(12:49):
happens six months, I have sixmonths to spare. If it doesn't
work out, it doesn't work out,I'll go find the next thing. But
two things happened during thosesix months. Number one, I
realized that Parag and I workreally closely together, and
that there was a massiveopportunity as we started to
(13:09):
talk to customers for what wewere trying to do. And the
second thing is, is that in thatsix months, when we fully
fleshed out the algorithmicpart, and the computational
parts of what Parag wasproposing, we realized that the
job was going to be a lot easierthan we had originally had
thought as well. And so oncethat was fully fleshed out, we
(13:31):
realized, boy, this is actuallynot just feasible, but like I
see a path to go and get thisdone. And that was when we
really started to raise moneyand really start to hire and go
after it.
Jonathan (13:43):
In this particular sort of problem area that Parag
identified as something thatwhere he could really move the
needle, my understanding of thisarea is that one of the big
limiters is the data set, likejust where do we get the data to
be able to then do all what I'llcall complexity analysis to be
(14:04):
able to identify certainproteins or derivatives that
could be great for drug target?
How are you addressing that lackof data?
Sujal Patel (14:13):
Yeah, if you take a step back for a second, I think
there are three... you know, theunderpinning your question is,
well, how do you deal with thetough stuff, and you said, data.
Data is one of the tough thingsthat we have to deal with.
There's a massive chunk of dataat the end of the pipeline,
there's a huge amount ofinformation that needs to be
(14:36):
gathered by instrumentation,which has never been built to
gather this much information.
And then on the biochemistryside, there's a huge amount of
complexity in trying to analyzethe molecules that we're going
after. And it might beinteresting just to kind of...
I'll double click quickly oneach of those areas, and then
you guys can delve in and tellme which areas you think are
most interesting. But in thebiochemistry side of it, one of
(14:58):
the things that made genomicsequencing possible was that
nature already has a mechanismto go and copy and read DNA. It
has to copy it for cell divisionso that each has a copy of the
DNA and has to read it so that aprocess can be undergone to
transform DNA into RNA, whichthen becomes protein. And in
(15:21):
order to make a genomicsequencer, Lumina, which is the
world leader by foreign genomicsequencing, built a system that
took a DNA fragment, amplifiedit, meaning copied it many, many
times and that enabled them toget signal amplification, which
made the process of measuring iteasier. For us, once something
(15:42):
becomes a protein in nature,there are no mechanisms to read
it back, or to copy it. And sowithout those mechanisms,
without a way to optically lookat it, because these objects are
miniscule, they're well belowthe optical resolution of any
microscope, because they'reorders of magnitude smaller than
(16:04):
the wavelength of light. Withouta way of measuring these
molecules, it's really,really... without a way to
borrow from nature in order tomeasure these molecules, it's
really difficult to develop amethod. And that's one of the
aha's that Parag had and thecrux of this method is that
instead of trying to make aspecific identification of what
a molecule is, if we borrow fromcomputer science, a technique
(16:28):
that probes the molecule, many,many, many times, each probe
leaking slightly differentinformation about the molecule
that you could computationallycombine that to get to a very
specific identity of what thatmolecule is. And then you have
to figure out how do you do thatin parallel for a large number
of molecules. That's the secondpart of the problem. In order to
(16:49):
make a dent in the world thatwe're going after, you have to
measure millions and billions ofmolecules in a single run of an
instrument. And that can't takemore than a few days. And the
reason you have to measure thatmuch is that your cells, each
cell has a lot of proteins init, roughly a million protein
molecules in every cell. In atypical pharmaceutical drug
(17:11):
development application, youmight be dealing with 96 well
plates, each one of them have1000 cells in it. So if you
can't turn through billions ofmolecules, you're not gonna want
to make a dent, even if youcould do better in terms of
protein identification, andwe've had to spend enormous
amount of effort figuring outhow can we, in our bio-chip have
density so that when you'rescanning it, that we're able to
(17:34):
get a lot of informationquickly, we've had to make a lot
of innovations in microscopy tofigure out how we can image
these molecules at a singlemolecule level. At speed, we had
to make a lot of innovations inthe microfluidic system, and all
the things that we need to probethese molecules over and over
(17:57):
again, so as a huge body of workthere. And then once that's
done, it goes into a computer, acomputer is dealing with 10, or
20 terabytes of raw informationcoming off the instrument every
day, that has to get reduced inreal time, these are images, so
they have to first get deskewed,they have to deal with
pincushion distortion, they haveto deal with all the fuzziness
(18:18):
of imaging at very low lightlevels, which is what we have to
deal with. And from there, thenwe reduce the dataset, we send
it to the cloud, and computationusing this algorithm that Parag
conceived of four years ago,requires hundreds of cores and
many hours of compute power. Andso that's just to get to
quantification of what's in mysample, what's in that drop of
(18:42):
blood. Then from there, thequestion is, well how do I use
that information to build betterdrugs, to build diagnostics to
personalized medicine? Andthat's a whole area of
exploration that we will be abig part of, but our customers
also be really integral pieceof.
Forrest Meyen (19:01):
Now, when you say each cell has millions of
different proteins, are youactually characterizing, you
know, a million proteins in asingle cell? Are you cataloging
everything?
Sujal Patel (19:15):
We believe, as a company that in order to be
useful to pharma, you need to beable to analyze the vast
majority of the proteins thatare in a very large number of
cells. So not just one cell, buthundreds or thousands of cells.
And so we're building technologyto be able to do that
effectively. One of the thingsthat's really interesting about
(19:38):
this company is that let me kindof draw an analogy back to the
genomic world in the genomicsworld. Before Illumina came
along. There was a method ofsequencing or a couple, but
Sanger sequencing being one ofthem, which was slow and
expensive. It was 100% rightor pretty much 100% right and
what Illumina figured out how todo is to make this a commodity:
(20:01):
how can I make it fast, cheap,and that's 99% accurate. For us,
there are some new companies inthe proteomics world that are
trying to do things likesequencing and trying to get to
100%, but those aren'ttechniques that are going to
lead to a massively parallelapproach that will democratize
(20:22):
access to the proteome for drugdevelopment companies for
pharma. We're building anapproach from day one that's
focused on we're not going toget 100% of the answer, but we'd
like to get some very highpercentage of the answer, but
we'd like to do it very quicklyat low cost. And so for us,
being able to analyze not justthe million proteins on average
(20:43):
that are in one cell, but beingable to do that for hundreds of
thousands is the goal of thiscompany.
Jonathan (20:48):
This topic is... and I mean it in the nicest ways, it's
the playground of PhDs who havestudied this for their academic
and professional lives, itseems. So what was it like for
you who had a very different andequally specialized, but a
different skill set, thenstarting to learn sort of the
(21:11):
what Parag has been working onfor so long?
Sujal Patel (21:14):
Yeah, it's been very interesting, right? So one
of the things that makes thecombination of Parag and I
unique is that, in order tobring this innovation to the
world, we're going to have tobuild an entirely new
instrument. That instrument haschemistry, it has biology, it
has imaging, biophysics,hardware, software. On our
(21:35):
staff, our mechanical engineers,software engineers, electrical
engineers, biophysicists workingside by side biochemists,
organic chemists, bioengineering majors, all of these
disciplines come together tobuild a complete solution. And
you know, Parag's experiencevirtually spans the entire
range, but I've spent the bulkof my career on the second part
(21:57):
of that, which is from thehardware engineering, to the
software and the data scienceside of things. And so the
combination is quite powerful.
And in addition to that, thedistribution model for products
in this world is very, verysimilar to the distribution
model that we had at Isilon, mylast company. Literally, you
just change the titles of thedifferent people in go-to-market
(22:17):
organization and they match upalmost identically in terms of
how they function. But for me, Ihave a lot to catch up on to get
going in this company. And itwas really, it was a fascinating
experience for me in the firstsix months to figure out how I
was going to get up to speed inthis world. So the first thing I
did was that I went on YouTube.
(22:42):
And I went and found everybiology and chemistry class that
I could and I set my speed to1.5x. And I started churning
through those as fast as Icould. And every day I
created... it was a list, Icalled it dumb questions of the
day. And I would make a list.
And at the end of the day, I'dcall Parag, and sometimes it'd
take 20 minutes, sometimes itwould take two hours, but we
(23:04):
would go through all of my dumbquestions. And Parag is saint.
He's an academic professor, he'svery patient. And he went
through all my questions. And Iwent through successive levels
of classes that were more andmore sophisticated to get to the
point where I had a basicunderstanding of the subject
matter that I was going into.
(23:26):
And then I started readingresearch papers, once I got
enough of that knowledge, and Iread probably 500 plus research
papers, maybe 1000 researchpapers in that first year. And I
still go through roughly one aday or something like that on
average. And then you start toreally consume that information.
And so today, I'm nowhere nearan expert in anything. But in
(23:48):
the one area where we arefocused, I've been getting
deeper and deeper and deeper.
Forrest Meyen (23:53):
That's amazing.
So complete immersion is yourstrategy.
Sujal Patel (23:56):
Yes, that's right.
Forrest Meyen (23:58):
That's got to be what you got to do. So I'm
curious about the... you call itan instrument, and it has all
these components and it seemsvery complex like is, is this
actually like a physical device?
Or is it more like a lab? Whatis the product?
Sujal Patel (24:16):
Yeah so ultimately, what we are building is an
instrument, much like a genomicsequencer genomic sequencers,
you can buy them from a numberof companies; they show up as a
box; the box sometimes would bethe size of a small bench. So
maybe it's a three foot wide bytwo foot by three foot high
instrument, depending on how bigit is and how fast it is. There
(24:40):
are some companies out therethat are building sequencers
that are more like the size of atoaster today, but it's a
physical piece of equipment. Andso for us, that's about the
first generation of sequencers;the next generation sequencing
wave, we're about the size ofkind of a few feet by a few feet
by five feet and that's what ourinitial product will be. It'll
be an instrument that's aboutthat size. And our job will be
(25:03):
to take that technology and makeit reliable and reproducible
enough that we can ship it toany lab in the world. And anyone
can easily run a proteomicanalysis. Now, between where we
are today, and that world, therewill certainly be a phase where
we've cracked the code, and wecan do it, we're just not ready
to give you a box yet. Andthat'll be a really interesting
(25:24):
phase of partnership between usa number of pharma organizations
and diagnostic companies. Andwe're starting to talk to many
of those partners today, as weare looking at starting some of
those early experiments nextyear.
Forrest Meyen (25:38):
And then I assume you'll also have the the
processing component, probablystill in the cloud. So the
instrument will just like, pluginto the wall and upload the
data?
Sujal Patel (25:49):
Yes, that is, right. So there's kind of
been... if you think about thedata pipeline, the instrument is
going to offload images to alocal computer, that computer is
going to have to receive massiveamount of information and reduce
it to something that's suitablefor sending over... the wider
your network. And then it'sgoing to have to go to the
(26:10):
cloud, the amount of computepower that we need is probably
infeasible for mostorganizations to have on site.
And so we'll send it up to thecloud for analysis. And then
we'll have a platform therewhere customers can access their
data, can look at their results,and then eventually, we'll
provide more and more analysiscapabilities in the cloud so
(26:31):
that our customers are receivinginsight, not just raw data.
Jonathan (26:35):
How would you compare Nautilus Biotechnology... draw a
distinction between that andsay, Seer Bio, where my
understanding is that they'resomewhere in this space as well.
Sujal Patel (26:54):
It'd be useful, I think, to discuss sort of the
broader landscape of thecompanies that are out there.
There are really two majorexisting categories of proteomic
analysis. One is, there's alarge amount of companies that
(27:16):
produce assays, which arespecific panels that will
identify the relative quantitiesof some known proteins in a
sample so if I take some cellsand I run an assay that can
identify 10 or 20 differentproteins; what the assay will
tell me is the relativeabundance of protein A is three
(27:37):
times as much as protein B, andI didn't see any protein C. It's
a relative abundance, becausethese assays are very fuzzy and
it only can support a very smallnumber of protein molecules,
because you have to haveantibodies for all the proteins
that you would want to analyzein an assay. And we as humanity,
haven't built more than a few1000 antibodies that will
(28:02):
identify different proteins. Andthe human proteome has 10 times
more, has 25,000 basic forms ofproteins to go after. And then
there's isoforms, andmodifications and other things.
So that's one category. So it'sa very targeted analysis. And
it's fuzzy because it's done inrelative abundance. The next
category is what's largely usedfor what's called discovery
(28:24):
proteomics, which is I have asample and I just want to know
the most about what's in thesample. And customers who want
to do the very best analysis usemass spectrometers today, and I
think you guys are familiar,because I know that you did a
recent video on massspectrometry, but what a mass
spectrometer is is a large,complicated instrument; it's
(28:46):
actually built for the atomicprogram, and its job is to weigh
fragments of molecules. And sowhat we do to use the mass
spectrometer for proteinanalysis is we take protein
molecules, we fragment thosemolecules into pieces, we send
them through the massspectrometer, and through a
complex process, it'sessentially telling you what the
(29:08):
atomic masses of each of thosefragments. And then on the other
side, we use of that set of verycomplicated bioinformatics to
reassemble that information intowhat the identities might have
been for the various proteinsthat I put in. This is a pretty
fuzzy process (29:24):
it suffers from two dramatic limitations. One is
that it has a very limiteddynamic range. And so if you
hand it blood, for example,without any treatment, most of
your blood is made up ofalbumin, which is basically just
a protein that's just there andis the bulk of your blood. If
(29:46):
you sent that to a massspectrometer, it would return
everything's albumin becauseit's always going to return the
most abundant stuff in a sample.
And so that's problem numberone. Problem number two is it's
not a very specific instrumenteither, and we're talking about
sensitivity and specificityhere. And so the thing that
customers have been doing isworking on techniques in front
(30:11):
to separate different types ofproteins from one another, so
that the mass spectrometer has abetter shot of returning useful
information. And typically whatpeople use is they use a UPLC,
which is essentially apurification method. So in one
incarnation of this system, itmay strip out from blood, the
(30:32):
top 14 proteins that show up sothat the mass spectrometer has a
shot of identifying someinteresting stuff beyond it.
Those techniques are onecomplicated, expensive; two,
they introduce bias into theresults; and three, they're just
not super effective. And so the8% number that I gave you is
roughly identifying 1500 to 2000proteins; that's using a very
(30:57):
complicated method of usingcolumns that go and remove the
abundant stuff from the bloodand make a better analysis. This
is really where Seer comes into
play (31:09):
Seer is a company that is focused on a better method of
preparing the sample upfrontthan what a UPLC can do. And so
that's a great method forimproving the efficacy of what
comes out of a massspectrometer, but it doesn't
fundamentally overcome thelimitations of a mass
spectrometer. And that's wherethis third group of companies
(31:30):
comes into play. And I thinkthat, in the third group, I
would put those companies likeus, which we don't know anyone
else like us, and then thosecompanies that are focused on
trying to sequence proteins.
Sequencing proteins is a hard
and complex challenge (31:44):
it requires very sophisticated
biochemistry, the methods arenot quite evolved enough where
you know it's really feasibletoday. And the other thing is
that you still have to fragmentthese proteins into peptides to
be able to sequence them,because there's limitations in
terms of the length that you cansequence. And so with that you
(32:05):
lose sensitivity a dramaticloss of sensitivity in the
analysis. So when we look at thedifferent competitors out there
in the marketplace and differentapproaches, that's kind of the
Jonathan (32:16):
So my understanding then is that the real value is
segmentation.
coming after hunting down theseless abundant proteins. And so,
connecting this back then towhat the general public... what
they need to know about thereally importance of this new
(32:39):
type of technology that you andyour team are building, it's
that if Nautilus Bio is able tohelp, say, a pharmaco and other
parts of its customer set to beable to identify less abundant
proteins, then that in turn, isvaluable, because then pharmacos
could develop better, moretargeted personalized medicine.
Sujal Patel (33:04):
That's correct.
Yes, you're exactly right. Sotoday, for a pharma
organization, let's you know, in
a simple model here (33:12):
if a pharma company has a set of samples
that have a disease cell, ordisease cells, and they have a
set of samples that have healthycells, what they want to do is
figure out what are thedifferences between these cells?
And once I figure out what thedifference is, and by the way,
most of the time, over 90% ofthe time, the difference is
(33:35):
there's a protein differencebetween these. The goal then is
how do I build a compound that'sgoing to be able to target that
protein difference? And then Ican deliver some kind of
therapeutic? So if that's whatI'm trying to do, I need to
ignore the abundant stuff,because that's the same stuff,
(33:56):
that's the stuff that is commonbetween diseased cells and
healthy cells. I need to focuson what are the rare things that
are in these cells that are verydifferent. We've been talking to
one of the top 10 pharams for along time, and the key scientist
there, one of the things thathe's told us is that in their
mass spectrometry core, the mostinteresting targets they find,
(34:20):
are at the very, very bottom ofthe detection threshold of mass
spec. So his comments us was ifyou could just push that down
5%. All that next stuff is goingto be super interesting, because
it's the rare stuff that we'relooking for. And so the ability
to get deeper, has dramaticimpacts on discovering what
(34:41):
targets these pharma companiesmight want to use for their next
therapeutic. And it also hasbroad implications to
understanding how thosetherapeutics will work, right?
Another example in pharma isthat when you've developed a
compound now you have to try tofigure out well what's the
therapeutic window? How much ofthis compound... how much of
this drug could I give somebodyin order to get the positive
(35:06):
effect that I want but not havenegative effects through the
body? And today measuringnegative effects through the
body... it's really just acrapshoot, right? I tried an
animal, I tried to human and wesee what happens. Because we
don't understand if I were toshow this compound to all the
other cells in your body, wehave no way of profiling in
(35:26):
advance what changes areoccurring, where's the cross
reactivity, the promise of beingable to dig further into your
cellular machinery in yourproteins is that we'll be able
to back up in the drugdevelopment process and
understand well, how would thisdrug impact my cardiac system,
my cells in my liver or mykidney? And if we could do that,
(35:47):
it would make the drugdevelopment process much more
efficient and much moreeffective.
Jonathan (35:53):
Walk me through then the interface between your
company and your customers? Isit that a customer comes saying,
hi Nautilus, we'd like, helpunderstanding an activation
pathway for a particulardisease. And then they give you
that sort of requirementsdocument and then Nautilus
(36:16):
returns and says, okay, alongthe pathway are all these
different types of proteins. Sotake a look at those?
Sujal Patel (36:24):
You ask a complex question. I think that there are
a broad range of business modelsthat we will employ as a
company. On one end, we willeventually have an instrument.
And we will sell to customersthe instrument, we'll sell
consumables that power it andwe'll sell them software as a
platform in the cloud that theycan use to analyze their
(36:46):
results. And in that model,they're largely in charge of
their innovation, and they'reusing our tools and our
platforms to be able to get tothat innovation. There are going
to be some customers that aregoing to want a closer
engagement with us, who aregoing to want help in trying to
(37:07):
identify for a particularpathway what's going on, they
are going to want to do thingsthat are more customized on our
platform than what our basicplatform can do. And in those
types of engagements, we aregoing to partner deeper with
them lending data sciencecapabilities and our knowledge
base to help them in theirmission. And some of those deals
(37:29):
may look more like a traditionalpartnership that a biotech will
often have with a pharma companywhere there will be milestone
payments, and maybe some upsideon the potential joint
development. And then the otherthing that we're thinking about
is that in the first few yearsof this new technology being
available, we're not going to beready to ship a box to a
(37:49):
customer and say, okay, it's allready to go, it's packaged up
neatly with a bow, just take itout of the box and run. And in
that early phase, all of ourengagements are going to be very
close to a development and makeengagements very close
collaborations. And so we're upfor helping our customers in
whatever way yields jointdiscoveries faster, because
(38:09):
that's in the benefit of bothour customer and us, right? Our
customer wants to get toinnovation so they can build
their next new drug, and we wantto go and prove out this new
platform and show that it'svaluable to the world because
it's going to be great for ourbusiness.
Forrest Meyen (38:22):
And just because it's like a hot topic right now
is are you doing anypartnerships, exploring
potential treatments or studyingCOVID-19?
Sujal Patel (38:35):
It's an interesting question. We're not far enough
along today to make a dent inthe COVID world. We've had many
conversations with customerswhere they're like, boy, I would
love to have this questionanswered, we're not there yet
today, right? We're notanalyzing customer samples, we
(38:55):
can't give you a huge proteomicanalysis but we'll see how fast
this pandemic will go away, butif I were to bet on it, I think
that our technology will get tothe point where it's valuable in
this pandemic at some pointbecause this isn't gonna go away
overnight.
Forrest Meyen (39:15):
And protecting against future pandemics as
well.
Sujal Patel (39:18):
For sure. That's a big part of what we want to make
sure we do here.
Jonathan (39:22):
I suppose when you're wearing your CEO hat, even
though you may not have theoffering on the table now to be
able to help sort of fight thisparticular fight, but it must be
awesome validation to havecustomers or prospective
customers come knocking lookingfor... that as you build out
(39:42):
these offerings that it'll bereally potentially game changing
for these folks. Rapiddevelopment of new medications
and so much more between that.
Sujal Patel (39:56):
Yeah, even more broadly than that, we're excited
with our customers' reaction,not just in COVID, but just
across a very broad range ofapplications. So I'll share with
you a funny anecdote. Werecently had a press release
that we hired this guy, NickNelson, to be our Chief Business
Officer. Well, Nick is a wellknown and super well regarded
(40:20):
business executive in thebiotech space. And he spent a
decade of his career atIllumina. And now that Nick has
been involved in some of ourpharma conversations and had
joint calls with Parag, myco-founder, one of the comments
that he made to Parag, and I'vetalked to probably about this in
(40:41):
the past was that, hey Parag,you realize, that these
conversations aren't supposed tobe 100% positive with customers.
You're supposed to have a one infive hit rate. One time, they
say, hey, can you do this, let'sgo do it at four to five.
They're like, hey, this isgreat, come back to me when it
works. And that's just not thecase here. Every customer
(41:01):
conversation is like, what can Ido tomorrow? How can I get to
the next step? How can you helpme in this, or this or this? And
it's really gratifying, right?
Because what it tells us is thatwe're really scratching an itch
that they have that's reallyimportant to their businesses,
and we're working as fast as wecan but we know that this
technology will be reallyimpactful when it gets out
(41:23):
there.
Forrest Meyen (41:25):
So what sort of things are you doing, given the
massive demand? And what are thethings that you're doing to work
as fast as you can, like whatsteps are you taking to kind of
accelerate your developmentprocess?
Sujal Patel (41:38):
Well, so that's an interesting question, right? We
are building a company for thelong run. So there is definitely
a balance there of speed, andmaking sure that we're building
a foundation that enables us toattract A players and build a
great corporate culture, andbuild an organization that's
(41:59):
going to be here in 20 years andbe a huge player in this future
proteomics industry. And so,there's a lot of tension to
building a company when you havethat kind of viewpoint, right? I
mean, first and foremost, if youwant to go and you want to move
quickly, you should have a lotof capital. And, you know, we've
(42:20):
been very fortunate that, youknow, just in the last three and
a half years, we've raised $109million of capital, we have a
blue-chip set of investorsacross the biotech and the tech
world, including somepublic-private crossover money.
And so we've got a great groupof investors that's given us the
capital to grow quickly, and togo after this problem
(42:43):
aggressively. But still, todaywe're still only a 55 person
company. I've told Parag and myrecruiting team, I'd like to be
at 100 like tomorrow, if Icould. But it's just not
feasible, right? We're trying tobuild an organization for the
long term. And we had thathealthy tension all the time of
well, I know I need some bodies,but they're not the right
(43:04):
people. And, we think we'remaking the right decisions for
the long term but that tensionis something I feel and I deal
with every day.
Jonathan (43:13):
With COVID-19, can I turn that question around then
on to as a startup founder andmanaging not just like you and
your co-founder's schedules, butnow you've got 55 people and
want to double that, how haveyou been navigating the
challenges of whether it'sshelter-in-place or you are
(43:36):
already sort of managing Iunderstand two locations in San
Carlos, California, Seattle,Washington what's been
working? And what may havefailed in that you sort of
figured out like, ah, thatdidn't work to make it better
for running a company in thisenvironment?
Sujal Patel (43:51):
Yeah, I think that this has been an area where we
have spent a lot of time andenergy and we've been pretty
intentional. And you have tokind of look at our two offices
separately, because all of ourwet science is done down in the
Bay Area. And in Seattle, it'slargely software engineering and
(44:12):
administrative functions. I'm uphere, my head of finance up
here. And those are things thatcan be done from home pretty
effectively. So kind of if youlook at the Bay Area, once we
got to the shelter-in-placeorder in California, we did what
all companies had to do, whichis we shut down and we assessed:
what do we need to do as anorganization to build a safe
(44:35):
environment where lab workerscan work once it's okay
according to our county and ourstate regulations? And for us,
that meant that we had to spenda lot of time energy and money
on improving our office so thatit could be safe in this COVID
world. We added UV sterilizersin our HVAC systems, which was a
(44:58):
huge amount of work andexpensive. We added plexiglass
and glass partitions betweenevery office in our open
cubicle area, in our labs. Wespaced things out, which made
our space utilization much lesseffective. And we created
policies to ensure mask wear.
And we're created policies toensure hygiene, we created a
(45:21):
schedule for electrostaticsterilization for our
facilities. Now, one thing afterthe otherandthat meant we were
able to get back to somewhateffective work in three weeks,
and then get back to prettyeffective work within the first
two months. Even today, we still
deal with issues (45:44):
childcare, people are on weird schedules
because of their own childcareissues. We are very conservative
with respect to employees ifthey're exposed, or even if
they're two degrees ofseparation exposed to COVID. And
(46:04):
so if your roommate's boyfriendhad COVID, we don't want you
back at work until you've hadsufficient tests that say that
you're negative. And so we'vetaken this conservative
approach, and I think what it'sdone is it's helped us to
maintain some goodwill with ouremployees. You know, up in
Seattle, we were closed for alonger period of time, but
(46:28):
eventually a number of us wantedto return back to the work
office, and today we let peopleup in Seattle work from wherever
they would like (46:36):
from their house is fine, or from the
office is fine. And so you know,at any given day, today, there's
four or five, I think there'sfive of us here, but we have
quite a bit of space relative tofive people. And so we're able
to physically distant, we havesome UV sterilizers. And we kind
of are smaller groups that wereable to kind of keep our own
(46:58):
pot. So that's how we've stayedsafe. You know, what I would
say, we could have done betterand I hear this from other CEOs
as well as is that this pandemicis... it's a really scary thing
for people, it's unnerving, it'screating stress. And I think
that, we tried to move veryquickly. And we probably could
(47:18):
be even more intentional withour communications, with our
discussions one on one withemployees about their concerns,
and make sure that they are allkind of following our journey
one step at a time. And I thinkwe did a great job. And I think
that that's always an area wherewe can improve, right?
Jonathan (47:37):
That's... yeah, that's absolutely important, is the
communication and being present,and so that you can really have
your hands on the issues andtrying to lead the ship, as
well. What are your prioritiesthen... looking about a year
out? What do you prioritize overothers for Nautilus Bio?
Sujal Patel (48:02):
Yeah, so for us, there are, broadly, a few things
that we need to do, and they'revery interrelated. We need to
get to the point over the courseof the next 12 to 18 months,
that we have the ability toroutinely take customer samples
(48:23):
in and return valuableinformation to them, that they
have no way to analyze and getthat information anywhere else
in the world. And to startbuilding the body of evidence
that our technique is reliableand returning correct results
reliably. That's job number one.
In order to do that, I have todouble the size of my SAP
overnight. It's an interestingphenomena... a year ago, year
(48:47):
and a half ago, you'd run anexperiment, you'd get the
results, you'd figure out whatto do next. We're not in
research mode largely anymore,we're in engineering mode,
optimizing the protocol,figuring out how to build the
next version of the thing thatwe have, figuring out how to
make this thing go 50% faster tomeet the spec that it needs. And
(49:08):
in that world, there's 100experiments in every single
person's bucket that's piled upthat needs to get done. And so
we just need to grow ourcapacity, and we've got to be
able to build our engineeringand our research and development
organizations to be able to getto the goal of going on the
(49:28):
product side. On the commercialside of our business, we're
really focused on refining ourbusiness models, widening our
engagement with customers andsigning the initial deals with
customers that enable us tostart to engage in some
meaningful areas that willdemonstrate the capabilities of
our platform and those arethat's kind of the stuff that is
largely keeping me busy rightnow. When I look forward, one of
(49:52):
the big goals for me over thecourse of the last year has been
to transform the company from aleadership perspective from an
early stage developmentorganization to a company that's
ready to go and introduce aproduct to the commercial
markets. And so we entered 2020,with Parag and myself being the
(50:13):
only two executives at thecompany. And over the course of
the year so far, we hired MaryGodwin, who runs the operations
organization for us. She's beenat this for decades and has
worked with me across a numberof companies. We hired Nick
Nelson, who I talked aboutearlier. And then I've already
hired two other executives thatare starting between now and the
(50:35):
end of the year. And so we'regoing into 2021 with a team
that's really set up to help usto move Nautilus into a
commercialization phase.
Forrest Meyen (50:46):
I had a quick question, before we get the
tying things up... so I wantedto ask you what's been the most
fun part of your experiencebuilding this company?
Sujal Patel (50:58):
What's been the most fun part of building this
company? So, I, I am a computerscientist at heart. I am an
engineer at heart, the most funfor me, has been actually
getting involved and solvingreal problems and doing real
engineering work. This company,my experience is actually quite
different than my last company.
And so when I found that Isilonin January of 2001, I was 26
(51:22):
years old, and no single thingabout being a CEO. And I had to
grow my staff from me to whatended up as 35 people at the end
of year one because we had amassive chunk of software to
tackle, to go and do what we'retrying to do at Isilon. Because
I didn't know how to do anythingas CEO that consumed all of my
(51:42):
time and that's what I did. Atthis company, Parag and I
started with just the twofounders for the first six or
seven months, working throughthe tough challenge of hey, is
that algorithm gonna work? And Itook on a very different role,
right? I mean, I've been a CEOfrom miniscule through public
(52:02):
company, that stuff took 10% ofmy time, it's super easy for me
to do. I spent 90% of my time onthe biology and chemistry,
education, and on coding, andactually working through the
algorithmic pieces. So Parag wasin the wet lab, six, seven
months, when it was just the twoof us, and I wasn't, I was in
front of a computer coding. Andit was a ton of fun. And believe
(52:25):
it or not, even two to two and ahalf years in the company's
life, I held on to doing some ofthe coding myself, which was a
refreshing change that I did notget to do at the last place.
Forrest Meyen (52:37):
So the coding skills didn't just fade away as
you're many years as CEO at theother company?
Sujal Patel (52:43):
They didn't fade away. I mean, I've done plenty
of side projects along the way,even as a CEO, but things those
things come back to you veryquickly. But now, I don't do any
more coding, because there's ahuge team of software engineers
here now. But I still getinvolved in a lot of the
engineering aspects of whatwe're building and how we're
(53:03):
going to move that forward. Andit's a really fun part of what I
do, right? I mean, there's a lotof things that are part of a
CEO's life that are mundane, orstressful or hard, or
people-oriented, or whateverthey might be. But for me, the
engineering stuff is reallyenergizing and fun.
Jonathan (53:23):
To conclude the episode, we'll be glad to give
you a space to share anythoughts or comments or
shameless plugs that you'd likeour audience to hear?
Sujal Patel (53:36):
Shameless plugs?
Well, look, I think I think thatI've probably given you enough
shameless plugs here. Hopefully,what I've been able to convey is
that I think this is a hugelytransformational company in the
making. And I am incrediblyexcited and passionate about it.
And I really appreciate you,Jonathan, Forrest, giving me the
(53:59):
opportunity to go and talk toyour viewers and tell you a
little bit more about what we'redoing. And I encourage you all
to stay tuned, because I thinkthat over the course of the next
year or two, this company isgoing to be announcing some
really interesting stuff that'sgoing to have real impact on
human health and humanhappiness. And that's, that's
why I'm in this.
Jonathan (54:20):
Oh, that's great to hear. So exciting and Nautilus
Biotechnology for the audienceis hiring. So if it looks like
you've got the skill set andchops, and working with Sujal
and his team, then it could be agreat home for you.
Sujal Patel (54:34):
I'm Sujal Patel, founder and CEO of Nautilus
Biotechnology. Stay tough!
Jonathan (54:39):
And that's a wrap of probing the proteome with Sujal
Patel of Nautilus Bio. If youare a loved one suffer from
allergies, our next episode maybe right up your alley. We sit
down with Conor Cullinane ofPirouette Medical. He and his
team are developing anautoinjector drug delivery
system. Their first offering isan epinephrine auto injector for
(55:00):
emergency treatment of severeallergic reactions. Subscribe
and join the mailing list so youstay in the loop. Meanwhile,
stay tough.
In order to make a dent in the world that we're
going after, you have to measuremillions and billions of
molecules in a single run of aninstrument, and that can't take
more than a few days, and thereason you have to measure that
much is that your cells, eachcell has a lot of proteins in
it.
Announcer (00:19):
Welcome to Tough Tech Today with Meyen and Miller.
This is the premier showfeaturing trailblazers who are
building technologies today tosolve tomorrow's toughest
challenges.
Jonathan (00:33):
Welcome to Tough Tech Today with Meyen and Miller.
Today, we have the honor ofspeaking with Sujal Patel,
co-founder of NautilusBiotechnology. Welcome, Sujal.
Sujal Patel (00:48):
Thank you, Jonathan and Forrest. Thanks for having
me.
Jonathan (00:50):
To get our audience up to speed with what you're
working on, could you tell us abit more about Nautilus
Biotechnology?
Sujal Patel (00:59):
Sure, Nautilus Biotechnology is a incredibly
interesting problem between thebiological domain, physical
sciences domain, and the techdomain. You know, if I'm to
explain the company in storyform, what I'd say is, is that
if you look at biotechnologyover the course of the last few
(01:20):
decades, and specifically welook at genomics, we as humanity
conquered genomics over thecourse of the last two decades
today, if I take a drop of yourblood, and I want to understand
what your genome is, I can getyou 99% of your genome for
$1,000. In a couple of days,it's a complete commodity. But
the issue is, is that yourgenome really doesn't change
(01:43):
from the day that you're born,the day that you die, it
actually is very static, itdoesn't contain anything real
time information about whetheryou're sick today or what's
going on. And because of that,there is this big move to try to
understand more of what's goingon in your cells, and your cells
are all made of proteins. And ifyou really want to understand
what's going on inside of humanbody, you have to get to the
(02:04):
protein level. And the issuethat exists in the world today
is that the very best techniquesthat we have to analyze proteins
from that same drop of blood orfrom any biological sample are
very different than genomics. Inthe genome, take a drop of
blood, $1,000, in a few days, Iget 99% of the answer. If I want
(02:25):
the proteome, which is themakeup of the proteins, I could
take that same drop of blood,the very best techniques on the
earth, leveraging complex piecesof equipment, like mass
spectrometers would take $50,000in a month to analyze that
sample. And at the conclusion ofthat, we would have identified
8% single digit 8% of theproteins that are the sample.
(02:47):
And so Nautilus was founded totry to improve that
dramatically.
Forrest Meyen (02:52):
So can you walk us through like exactly why we
would want to know what's in theparticular proteins? Because I
imagine your DNA encodes for allof those proteins to allow them
to exist? Are you trying to getinformation on environmental
factors or different things fromthose tests?
Sujal Patel (03:14):
It's much more than just environmental factors. But
I think that Forrest you're onthe right track there. So
obviously, your genome encodesall of the possible proteins
that could be created, but for awide variety of reasons, what's
occurring in your cells, and theproteins in your cells do not
match in any particular way whatyour genomic if we look at it
(03:37):
in software terms, what yoursoftware code says. And that's
the reason why if you took twotwins, and you let them age to
their 40, and one took care ofthemselves and one ate potato
chips all day and neverexercised, they'd look
completely different. Theenvironmental factors, the
health factors, what's going onin that particular body, all
(03:57):
those things are encoded in yourproteins. This is the reason why
if you looked at the FDAapproved drugs, over 90% of the
FDA approved drugs targetproteins. Most of our
diagnostics target proteins. Andone of the side effects of this
inability to measure proteinseffectively is that drug
(04:19):
development has gotten lessefficient, and it's gotten
worse. So over the last twodecades, if you looked at the
number of new FDA approved drugsevery year, it's basically dead
flat. And that's in the face ofa quadrupling of the global R&D
spend in pharma from about 50billion to 200 billion so we've
(04:39):
quadrupled spend, but FDAapprovals is flat. Because of
that, only two out of 12marketed drugs that reach the
market actually return the R&Ddollars that went into the
advocacy of the timelines havebecome worse and extended. And
not understanding the proteinswhich make up all of your cells
(05:00):
is a huge part of that problemthat's that's been created here.
Jonathan (05:03):
This is a... really complicated problem set. And so
I'm curious with your backgroundwith as CEO of Isilon Systems
which from my understanding wasfar and away from a biotech
company. And so could you tellus a bit... how to connect those
(05:28):
dots between your work withIsilon and then now find
yourself at the helm of a reallyexciting biotech company.
Sujal Patel (05:36):
Yeah, it's a great question, Jonathan. Because the
path for me to get from my lastcompany to here is a really
interesting path. And it's oneof those sort of small world
types of stories. So my lastcompany... well, let me kind of
back up. Up until four yearsago, my entire career has been
in the tech world. And the lastcompany that I spoke to was
(06:00):
company called Isilon, which wasa company founded in January of
2001, went public in 2006. Andthen we eventually sold it for
2.6 billion in 2010. Thatcompany was focused on building
a new storage architecture forunstructured information, things
like digital content, video andimages machine generated data.
(06:21):
Two decades ago, storage systemsof the time, were really focused
on string text based informationand databases and credit card
transactions. And so we built ascalable architecture. And we
took that to the world focusedon different verticals that were
undergoing a digitaltransformation. So we started
selling into companies doingphoto sharing on the internet,
(06:42):
into film broadcasts andtelevision which was undergoing
a transition from analog todigital. And then we moved into
manufacturing and semiconductorand by '04 and '05, life
sciences and biomedical researchbecame a really big market for
us. And this revolution thatoccurred in genetic sequencing
was a big part of the successthat we had in that vertical
(07:03):
market. Back in 2004, which is16 years ago, I met a guy named
Parag Malik and Parag became avery large customer of mine at
Isilon supporting his proteomicsResearch Lab at Cedars Sinai
Medical Center, and Parag 16years later, is my co-founder
(07:23):
here at Nautilus. And that, ourpaths intersected many times
over the course of those 16years. First, Parag was in the
role of a customer to me, and Igot to know Prague really well
and thought he was one of thesmartest guys that I ever met.
About nine years ago, Parag lefthis role at Cedars Sinai and
(07:47):
went to Stanford to teach andstart a research lab, which was
really building technologies topursue a personalized medicine,
with a focus of sitting at theintersection of tech and biotech
products, kind of a uniqueanimal. And he has academic
degrees in both biochemistry andcomputer science. And my wife
and I were so impressed with thework that he was trying to do in
(08:10):
his lab and then he has donethat we decided to
philanthropically support hislab. And we've done that for the
last nine years, which built areally close relationship
between Parag and I, and thatled to Parag calling me up in
2016 and saying, Hey, I'm gonnastart a company because I have
an idea and let me tell youabout it. And what started as a
(08:31):
conversation, soliciting advicevery quickly turned into two
founders going off and trying tosolve a really tough, really
hard problem. And it's been anamazing ride for four years of
kind of moving from the crackpotidea phase through feasible and
now into real engineering.
Jonathan (08:50):
That's so cool to hear how it's like an organic
relationship that developed,that you found that you could...
worked well together because ofgo founding a company is
non-trivial, and absolutely astrain on a relationship. What
were you doing when you got thatcall in 2016? And were you
(09:11):
looking for your next bigcompany to start?
Sujal Patel (09:15):
Yeah, so it's interesting, right. So we sold
Isilon in December of 2010. Andwe just closed our 100 million
dollar quarter. And I hadcommitted to the acquiring
company that I would grow thebusiness to a billion dollar run
rate and then I would leave. Wewere in a very high growth
rate that only took 23 months,and then I left. And I did some
(09:36):
consulting for them to help themthrough a few other things that
were going on. But largely, Istarted to think about what I
wanted to do next. And I hungout with venture capitalists, I
made a lot of investments incompanies that total I've made
about 80 investments in the last10 years in private companies. I
had six boards at some point andI started to work with
entrepreneurs and about sixmonths in I realized that I
(10:03):
couldn't be on boards andinvest; I really was too young,
too passionate, too excited togo and actually build something.
And I knew I wanted to go outand do something again. But I
also knew that I didn't want todo enterprise software,
enterprise hardware, I didn'twant to do something that was
like what I did before. And Ialso had this burning desire to
go work on really hard problemsthat have big societal impact.
(10:26):
And so I was only really lookingin three areas I was looking in
healthcare technology, I waslooking in clean energy, and I
was looking at a couple of ideasin consumer, which would have
very broad implications. Andthat's it. And none of those
ideas ever graduated to thepoint where I said, that was the
one for me until Parag came andit literally was that first one
(10:48):
hour call with Parag, I wentthrough it, I'm like, okay, if
anyone were to bring me an idealike this, I'd call Parag and
say, Parag, what do you thinkabout this idea? And then I said
to myself, Parag is going to goput his Stanford career which is
really successful on hold, ifthis was an incredibly exciting
in the opportunity of alifetime. And I got there in one
(11:10):
hour, obviously, there are a lotmore conversations from there,
but in one hour, I was like,Okay, this is it like this is
that thing. And it's been it'sbeen super exciting, getting
into a whole new and differentspace.
Forrest Meyen (11:23):
So you called the early phase of the company, or
the crackpot idea phase. Now wasit really a crackpot idea? Did
you shift a lot into what it isnow? Or were you just getting
more information to gainconfidence that you had
something really special?
Sujal Patel (11:43):
So that's an interesting question, right? I
mean, tongue in cheek, we callthis the crackpot idea phase in
2016. But what I will tell youis that the idea was really
unformed at that point. But wehad an idea that the method that
Parag was proposing to analyzehuge numbers of proteins and get
(12:04):
to a very specificidentification with those
proteins. And we an idea thatthat was possible, but there
were a huge number of unansweredquestions. And a big chunk of it
was related to the corealgorithm that Parag conceived
of that is the underpinnings ofwhat we do. And a big part of it
is how do we scale this up anddo it reliably at massive scale
(12:28):
at low cost. And now, that part,is really the bulk of what we
took the last four yearsworking... what we've done over
the last four years. The firstsix months were algorithmically
moving it from crackpot idea to,okay, this really can happen.
And in all honesty, I said tomyself, well, let's go see what
(12:49):
happens six months, I have sixmonths to spare. If it doesn't
work out, it doesn't work out,I'll go find the next thing. But
two things happened during thosesix months. Number one, I
realized that Parag and I workreally closely together, and
that there was a massiveopportunity as we started to
(13:09):
talk to customers for what wewere trying to do. And the
second thing is, is that in thatsix months, when we fully
fleshed out the algorithmicpart, and the computational
parts of what Parag wasproposing, we realized that the
job was going to be a lot easierthan we had originally had
thought as well. And so oncethat was fully fleshed out, we
(13:31):
realized, boy, this is actuallynot just feasible, but like I
see a path to go and get thisdone. And that was when we
really started to raise moneyand really start to hire and go
after it.
Jonathan (13:43):
In this particular sort of problem area that Parag
identified as something thatwhere he could really move the
needle, my understanding of thisarea is that one of the big
limiters is the data set, likejust where do we get the data to
be able to then do all what I'llcall complexity analysis to be
(14:04):
able to identify certainproteins or derivatives that
could be great for drug target?
How are you addressing that lackof data?
Sujal Patel (14:13):
Yeah, if you take a step back for a second, I think
there are three... you know, theunderpinning your question is,
well, how do you deal with thetough stuff, and you said, data.
Data is one of the tough thingsthat we have to deal with.
There's a massive chunk of dataat the end of the pipeline,
there's a huge amount ofinformation that needs to be
(14:36):
gathered by instrumentation,which has never been built to
gather this much information.
And then on the biochemistryside, there's a huge amount of
complexity in trying to analyzethe molecules that we're going
after. And it might beinteresting just to kind of...
I'll double click quickly oneach of those areas, and then
you guys can delve in and tellme which areas you think are
most interesting. But in thebiochemistry side of it, one of
(14:58):
the things that made genomicsequencing possible was that
nature already has a mechanismto go and copy and read DNA. It
has to copy it for cell divisionso that each has a copy of the
DNA and has to read it so that aprocess can be undergone to
transform DNA into RNA, whichthen becomes protein. And in
(15:21):
order to make a genomicsequencer, Lumina, which is the
world leader by foreign genomicsequencing, built a system that
took a DNA fragment, amplifiedit, meaning copied it many, many
times and that enabled them toget signal amplification, which
made the process of measuring iteasier. For us, once something
(15:42):
becomes a protein in nature,there are no mechanisms to read
it back, or to copy it. And sowithout those mechanisms,
without a way to optically lookat it, because these objects are
miniscule, they're well belowthe optical resolution of any
microscope, because they'reorders of magnitude smaller than
(16:04):
the wavelength of light. Withouta way of measuring these
molecules, it's really,really... without a way to
borrow from nature in order tomeasure these molecules, it's
really difficult to develop amethod. And that's one of the
aha's that Parag had and thecrux of this method is that
instead of trying to make aspecific identification of what
a molecule is, if we borrow fromcomputer science, a technique
(16:28):
that probes the molecule, many,many, many times, each probe
leaking slightly differentinformation about the molecule
that you could computationallycombine that to get to a very
specific identity of what thatmolecule is. And then you have
to figure out how do you do thatin parallel for a large number
of molecules. That's the secondpart of the problem. In order to
(16:49):
make a dent in the world thatwe're going after, you have to
measure millions and billions ofmolecules in a single run of an
instrument. And that can't takemore than a few days. And the
reason you have to measure thatmuch is that your cells, each
cell has a lot of proteins init, roughly a million protein
molecules in every cell. In atypical pharmaceutical drug
(17:11):
development application, youmight be dealing with 96 well
plates, each one of them have1000 cells in it. So if you
can't turn through billions ofmolecules, you're not gonna want
to make a dent, even if youcould do better in terms of
protein identification, andwe've had to spend enormous
amount of effort figuring outhow can we, in our bio-chip have
density so that when you'rescanning it, that we're able to
(17:34):
get a lot of informationquickly, we've had to make a lot
of innovations in microscopy tofigure out how we can image
these molecules at a singlemolecule level. At speed, we had
to make a lot of innovations inthe microfluidic system, and all
the things that we need to probethese molecules over and over
(17:57):
again, so as a huge body of workthere. And then once that's
done, it goes into a computer, acomputer is dealing with 10, or
20 terabytes of raw informationcoming off the instrument every
day, that has to get reduced inreal time, these are images, so
they have to first get deskewed,they have to deal with
pincushion distortion, they haveto deal with all the fuzziness
(18:18):
of imaging at very low lightlevels, which is what we have to
deal with. And from there, thenwe reduce the dataset, we send
it to the cloud, and computationusing this algorithm that Parag
conceived of four years ago,requires hundreds of cores and
many hours of compute power. Andso that's just to get to
quantification of what's in mysample, what's in that drop of
(18:42):
blood. Then from there, thequestion is, well how do I use
that information to build betterdrugs, to build diagnostics to
personalized medicine? Andthat's a whole area of
exploration that we will be abig part of, but our customers
also be really integral pieceof.
Forrest Meyen (19:01):
Now, when you say each cell has millions of
different proteins, are youactually characterizing, you
know, a million proteins in asingle cell? Are you cataloging
everything?
Sujal Patel (19:15):
We believe, as a company that in order to be
useful to pharma, you need to beable to analyze the vast
majority of the proteins thatare in a very large number of
cells. So not just one cell, buthundreds or thousands of cells.
And so we're building technologyto be able to do that
effectively. One of the thingsthat's really interesting about
(19:38):
this company is that let me kindof draw an analogy back to the
genomic world in the genomicsworld. Before Illumina came
along. There was a method ofsequencing or a couple, but
Sanger sequencing being one ofthem, which was slow and
expensive. It was 100% rightor pretty much 100% right and
what Illumina figured out how todo is to make this a commodity:
(20:01):
how can I make it fast, cheap,and that's 99% accurate. For us,
there are some new companies inthe proteomics world that are
trying to do things likesequencing and trying to get to
100%, but those aren'ttechniques that are going to
lead to a massively parallelapproach that will democratize
(20:22):
access to the proteome for drugdevelopment companies for
pharma. We're building anapproach from day one that's
focused on we're not going toget 100% of the answer, but we'd
like to get some very highpercentage of the answer, but
we'd like to do it very quicklyat low cost. And so for us,
being able to analyze not justthe million proteins on average
(20:43):
that are in one cell, but beingable to do that for hundreds of
thousands is the goal of thiscompany.
Jonathan (20:48):
This topic is... and I mean it in the nicest ways, it's
the playground of PhDs who havestudied this for their academic
and professional lives, itseems. So what was it like for
you who had a very different andequally specialized, but a
different skill set, thenstarting to learn sort of the
(21:11):
what Parag has been working onfor so long?
Sujal Patel (21:14):
Yeah, it's been very interesting, right? So one
of the things that makes thecombination of Parag and I
unique is that, in order tobring this innovation to the
world, we're going to have tobuild an entirely new
instrument. That instrument haschemistry, it has biology, it
has imaging, biophysics,hardware, software. On our
(21:35):
staff, our mechanical engineers,software engineers, electrical
engineers, biophysicists workingside by side biochemists,
organic chemists, bioengineering majors, all of these
disciplines come together tobuild a complete solution. And
you know, Parag's experiencevirtually spans the entire
range, but I've spent the bulkof my career on the second part
(21:57):
of that, which is from thehardware engineering, to the
software and the data scienceside of things. And so the
combination is quite powerful.
And in addition to that, thedistribution model for products
in this world is very, verysimilar to the distribution
model that we had at Isilon, mylast company. Literally, you
just change the titles of thedifferent people in go-to-market
(22:17):
organization and they match upalmost identically in terms of
how they function. But for me, Ihave a lot to catch up on to get
going in this company. And itwas really, it was a fascinating
experience for me in the firstsix months to figure out how I
was going to get up to speed inthis world. So the first thing I
did was that I went on YouTube.
(22:42):
And I went and found everybiology and chemistry class that
I could and I set my speed to1.5x. And I started churning
through those as fast as Icould. And every day I
created... it was a list, Icalled it dumb questions of the
day. And I would make a list.
And at the end of the day, I'dcall Parag, and sometimes it'd
take 20 minutes, sometimes itwould take two hours, but we
(23:04):
would go through all of my dumbquestions. And Parag is saint.
He's an academic professor, he'svery patient. And he went
through all my questions. And Iwent through successive levels
of classes that were more andmore sophisticated to get to the
point where I had a basicunderstanding of the subject
matter that I was going into.
(23:26):
And then I started readingresearch papers, once I got
enough of that knowledge, and Iread probably 500 plus research
papers, maybe 1000 researchpapers in that first year. And I
still go through roughly one aday or something like that on
average. And then you start toreally consume that information.
And so today, I'm nowhere nearan expert in anything. But in
(23:48):
the one area where we arefocused, I've been getting
deeper and deeper and deeper.
Forrest Meyen (23:53):
That's amazing.
So complete immersion is yourstrategy.
Sujal Patel (23:56):
Yes, that's right.
Forrest Meyen (23:58):
That's got to be what you got to do. So I'm
curious about the... you call itan instrument, and it has all
these components and it seemsvery complex like is, is this
actually like a physical device?
Or is it more like a lab? Whatis the product?
Sujal Patel (24:16):
Yeah so ultimately, what we are building is an
instrument, much like a genomicsequencer genomic sequencers,
you can buy them from a numberof companies; they show up as a
box; the box sometimes would bethe size of a small bench. So
maybe it's a three foot wide bytwo foot by three foot high
instrument, depending on how bigit is and how fast it is. There
(24:40):
are some companies out therethat are building sequencers
that are more like the size of atoaster today, but it's a
physical piece of equipment. Andso for us, that's about the
first generation of sequencers;the next generation sequencing
wave, we're about the size ofkind of a few feet by a few feet
by five feet and that's what ourinitial product will be. It'll
be an instrument that's aboutthat size. And our job will be
(25:03):
to take that technology and makeit reliable and reproducible
enough that we can ship it toany lab in the world. And anyone
can easily run a proteomicanalysis. Now, between where we
are today, and that world, therewill certainly be a phase where
we've cracked the code, and wecan do it, we're just not ready
to give you a box yet. Andthat'll be a really interesting
(25:24):
phase of partnership between usa number of pharma organizations
and diagnostic companies. Andwe're starting to talk to many
of those partners today, as weare looking at starting some of
those early experiments nextyear.
Forrest Meyen (25:38):
And then I assume you'll also have the the
processing component, probablystill in the cloud. So the
instrument will just like, pluginto the wall and upload the
data?
Sujal Patel (25:49):
Yes, that is, right. So there's kind of
been... if you think about thedata pipeline, the instrument is
going to offload images to alocal computer, that computer is
going to have to receive massiveamount of information and reduce
it to something that's suitablefor sending over... the wider
your network. And then it'sgoing to have to go to the
(26:10):
cloud, the amount of computepower that we need is probably
infeasible for mostorganizations to have on site.
And so we'll send it up to thecloud for analysis. And then
we'll have a platform therewhere customers can access their
data, can look at their results,and then eventually, we'll
provide more and more analysiscapabilities in the cloud so
(26:31):
that our customers are receivinginsight, not just raw data.
Jonathan (26:35):
How would you compare Nautilus Biotechnology... draw a
distinction between that andsay, Seer Bio, where my
understanding is that they'resomewhere in this space as well.
Sujal Patel (26:54):
It'd be useful, I think, to discuss sort of the
broader landscape of thecompanies that are out there.
There are really two majorexisting categories of proteomic
analysis. One is, there's alarge amount of companies that
(27:16):
produce assays, which arespecific panels that will
identify the relative quantitiesof some known proteins in a
sample so if I take some cellsand I run an assay that can
identify 10 or 20 differentproteins; what the assay will
tell me is the relativeabundance of protein A is three
(27:37):
times as much as protein B, andI didn't see any protein C. It's
a relative abundance, becausethese assays are very fuzzy and
it only can support a very smallnumber of protein molecules,
because you have to haveantibodies for all the proteins
that you would want to analyzein an assay. And we as humanity,
haven't built more than a few1000 antibodies that will
(28:02):
identify different proteins. Andthe human proteome has 10 times
more, has 25,000 basic forms ofproteins to go after. And then
there's isoforms, andmodifications and other things.
So that's one category. So it'sa very targeted analysis. And
it's fuzzy because it's done inrelative abundance. The next
category is what's largely usedfor what's called discovery
(28:24):
proteomics, which is I have asample and I just want to know
the most about what's in thesample. And customers who want
to do the very best analysis usemass spectrometers today, and I
think you guys are familiar,because I know that you did a
recent video on massspectrometry, but what a mass
spectrometer is is a large,complicated instrument; it's
(28:46):
actually built for the atomicprogram, and its job is to weigh
fragments of molecules. And sowhat we do to use the mass
spectrometer for proteinanalysis is we take protein
molecules, we fragment thosemolecules into pieces, we send
them through the massspectrometer, and through a
complex process, it'sessentially telling you what the
(29:08):
atomic masses of each of thosefragments. And then on the other
side, we use of that set of verycomplicated bioinformatics to
reassemble that information intowhat the identities might have
been for the various proteinsthat I put in. This is a pretty
fuzzy process (29:24):
it suffers from two dramatic limitations. One is
that it has a very limiteddynamic range. And so if you
hand it blood, for example,without any treatment, most of
your blood is made up ofalbumin, which is basically just
a protein that's just there andis the bulk of your blood. If
(29:46):
you sent that to a massspectrometer, it would return
everything's albumin becauseit's always going to return the
most abundant stuff in a sample.
And so that's problem numberone. Problem number two is it's
not a very specific instrumenteither, and we're talking about
sensitivity and specificityhere. And so the thing that
customers have been doing isworking on techniques in front
(30:11):
to separate different types ofproteins from one another, so
that the mass spectrometer has abetter shot of returning useful
information. And typically whatpeople use is they use a UPLC,
which is essentially apurification method. So in one
incarnation of this system, itmay strip out from blood, the
(30:32):
top 14 proteins that show up sothat the mass spectrometer has a
shot of identifying someinteresting stuff beyond it.
Those techniques are onecomplicated, expensive; two,
they introduce bias into theresults; and three, they're just
not super effective. And so the8% number that I gave you is
roughly identifying 1500 to 2000proteins; that's using a very
(30:57):
complicated method of usingcolumns that go and remove the
abundant stuff from the bloodand make a better analysis. This
is really where Seer comes into
play (31:09):
Seer is a company that is focused on a better method of
preparing the sample upfrontthan what a UPLC can do. And so
that's a great method forimproving the efficacy of what
comes out of a massspectrometer, but it doesn't
fundamentally overcome thelimitations of a mass
spectrometer. And that's wherethis third group of companies
(31:30):
comes into play. And I thinkthat, in the third group, I
would put those companies likeus, which we don't know anyone
else like us, and then thosecompanies that are focused on
trying to sequence proteins.
Sequencing proteins is a hard
and complex challenge (31:44):
it requires very sophisticated
biochemistry, the methods arenot quite evolved enough where
you know it's really feasibletoday. And the other thing is
that you still have to fragmentthese proteins into peptides to
be able to sequence them,because there's limitations in
terms of the length that you cansequence. And so with that you
(32:05):
lose sensitivity a dramaticloss of sensitivity in the
analysis. So when we look at thedifferent competitors out there
in the marketplace and differentapproaches, that's kind of the
Jonathan (32:16):
So my understanding then is that the real value is
segmentation.
coming after hunting down theseless abundant proteins. And so,
connecting this back then towhat the general public... what
they need to know about thereally importance of this new
(32:39):
type of technology that you andyour team are building, it's
that if Nautilus Bio is able tohelp, say, a pharmaco and other
parts of its customer set to beable to identify less abundant
proteins, then that in turn, isvaluable, because then pharmacos
could develop better, moretargeted personalized medicine.
Sujal Patel (33:04):
That's correct.
Yes, you're exactly right. Sotoday, for a pharma
organization, let's you know, in
a simple model here (33:12):
if a pharma company has a set of samples
that have a disease cell, ordisease cells, and they have a
set of samples that have healthycells, what they want to do is
figure out what are thedifferences between these cells?
And once I figure out what thedifference is, and by the way,
most of the time, over 90% ofthe time, the difference is
(33:35):
there's a protein differencebetween these. The goal then is
how do I build a compound that'sgoing to be able to target that
protein difference? And then Ican deliver some kind of
therapeutic? So if that's whatI'm trying to do, I need to
ignore the abundant stuff,because that's the same stuff,
(33:56):
that's the stuff that is commonbetween diseased cells and
healthy cells. I need to focuson what are the rare things that
are in these cells that are verydifferent. We've been talking to
one of the top 10 pharams for along time, and the key scientist
there, one of the things thathe's told us is that in their
mass spectrometry core, the mostinteresting targets they find,
(34:20):
are at the very, very bottom ofthe detection threshold of mass
spec. So his comments us was ifyou could just push that down
5%. All that next stuff is goingto be super interesting, because
it's the rare stuff that we'relooking for. And so the ability
to get deeper, has dramaticimpacts on discovering what
(34:41):
targets these pharma companiesmight want to use for their next
therapeutic. And it also hasbroad implications to
understanding how thosetherapeutics will work, right?
Another example in pharma isthat when you've developed a
compound now you have to try tofigure out well what's the
therapeutic window? How much ofthis compound... how much of
this drug could I give somebodyin order to get the positive
(35:06):
effect that I want but not havenegative effects through the
body? And today measuringnegative effects through the
body... it's really just acrapshoot, right? I tried an
animal, I tried to human and wesee what happens. Because we
don't understand if I were toshow this compound to all the
other cells in your body, wehave no way of profiling in
(35:26):
advance what changes areoccurring, where's the cross
reactivity, the promise of beingable to dig further into your
cellular machinery in yourproteins is that we'll be able
to back up in the drugdevelopment process and
understand well, how would thisdrug impact my cardiac system,
my cells in my liver or mykidney? And if we could do that,
(35:47):
it would make the drugdevelopment process much more
efficient and much moreeffective.
Jonathan (35:53):
Walk me through then the interface between your
company and your customers? Isit that a customer comes saying,
hi Nautilus, we'd like, helpunderstanding an activation
pathway for a particulardisease. And then they give you
that sort of requirementsdocument and then Nautilus
(36:16):
returns and says, okay, alongthe pathway are all these
different types of proteins. Sotake a look at those?
Sujal Patel (36:24):
You ask a complex question. I think that there are
a broad range of business modelsthat we will employ as a
company. On one end, we willeventually have an instrument.
And we will sell to customersthe instrument, we'll sell
consumables that power it andwe'll sell them software as a
platform in the cloud that theycan use to analyze their
(36:46):
results. And in that model,they're largely in charge of
their innovation, and they'reusing our tools and our
platforms to be able to get tothat innovation. There are going
to be some customers that aregoing to want a closer
engagement with us, who aregoing to want help in trying to
(37:07):
identify for a particularpathway what's going on, they
are going to want to do thingsthat are more customized on our
platform than what our basicplatform can do. And in those
types of engagements, we aregoing to partner deeper with
them lending data sciencecapabilities and our knowledge
base to help them in theirmission. And some of those deals
(37:29):
may look more like a traditionalpartnership that a biotech will
often have with a pharma companywhere there will be milestone
payments, and maybe some upsideon the potential joint
development. And then the otherthing that we're thinking about
is that in the first few yearsof this new technology being
available, we're not going to beready to ship a box to a
(37:49):
customer and say, okay, it's allready to go, it's packaged up
neatly with a bow, just take itout of the box and run. And in
that early phase, all of ourengagements are going to be very
close to a development and makeengagements very close
collaborations. And so we're upfor helping our customers in
whatever way yields jointdiscoveries faster, because
(38:09):
that's in the benefit of bothour customer and us, right? Our
customer wants to get toinnovation so they can build
their next new drug, and we wantto go and prove out this new
platform and show that it'svaluable to the world because
it's going to be great for ourbusiness.
Forrest Meyen (38:22):
And just because it's like a hot topic right now
is are you doing anypartnerships, exploring
potential treatments or studyingCOVID-19?
Sujal Patel (38:35):
It's an interesting question. We're not far enough
along today to make a dent inthe COVID world. We've had many
conversations with customerswhere they're like, boy, I would
love to have this questionanswered, we're not there yet
today, right? We're notanalyzing customer samples, we
(38:55):
can't give you a huge proteomicanalysis but we'll see how fast
this pandemic will go away, butif I were to bet on it, I think
that our technology will get tothe point where it's valuable in
this pandemic at some pointbecause this isn't gonna go away
overnight.
Forrest Meyen (39:15):
And protecting against future pandemics as
well.
Sujal Patel (39:18):
For sure. That's a big part of what we want to make
sure we do here.
Jonathan (39:22):
I suppose when you're wearing your CEO hat, even
though you may not have theoffering on the table now to be
able to help sort of fight thisparticular fight, but it must be
awesome validation to havecustomers or prospective
customers come knocking lookingfor... that as you build out
(39:42):
these offerings that it'll bereally potentially game changing
for these folks. Rapiddevelopment of new medications
and so much more between that.
Sujal Patel (39:56):
Yeah, even more broadly than that, we're excited
with our customers' reaction,not just in COVID, but just
across a very broad range ofapplications. So I'll share with
you a funny anecdote. Werecently had a press release
that we hired this guy, NickNelson, to be our Chief Business
Officer. Well, Nick is a wellknown and super well regarded
(40:20):
business executive in thebiotech space. And he spent a
decade of his career atIllumina. And now that Nick has
been involved in some of ourpharma conversations and had
joint calls with Parag, myco-founder, one of the comments
that he made to Parag, and I'vetalked to probably about this in
(40:41):
the past was that, hey Parag,you realize, that these
conversations aren't supposed tobe 100% positive with customers.
You're supposed to have a one infive hit rate. One time, they
say, hey, can you do this, let'sgo do it at four to five.
They're like, hey, this isgreat, come back to me when it
works. And that's just not thecase here. Every customer
(41:01):
conversation is like, what can Ido tomorrow? How can I get to
the next step? How can you helpme in this, or this or this? And
it's really gratifying, right?
Because what it tells us is thatwe're really scratching an itch
that they have that's reallyimportant to their businesses,
and we're working as fast as wecan but we know that this
technology will be reallyimpactful when it gets out
(41:23):
there.
Forrest Meyen (41:25):
So what sort of things are you doing, given the
massive demand? And what are thethings that you're doing to work
as fast as you can, like whatsteps are you taking to kind of
accelerate your developmentprocess?
Sujal Patel (41:38):
Well, so that's an interesting question, right? We
are building a company for thelong run. So there is definitely
a balance there of speed, andmaking sure that we're building
a foundation that enables us toattract A players and build a
great corporate culture, andbuild an organization that's
(41:59):
going to be here in 20 years andbe a huge player in this future
proteomics industry. And so,there's a lot of tension to
building a company when you havethat kind of viewpoint, right? I
mean, first and foremost, if youwant to go and you want to move
quickly, you should have a lotof capital. And, you know, we've
(42:20):
been very fortunate that, youknow, just in the last three and
a half years, we've raised $109million of capital, we have a
blue-chip set of investorsacross the biotech and the tech
world, including somepublic-private crossover money.
And so we've got a great groupof investors that's given us the
capital to grow quickly, and togo after this problem
(42:43):
aggressively. But still, todaywe're still only a 55 person
company. I've told Parag and myrecruiting team, I'd like to be
at 100 like tomorrow, if Icould. But it's just not
feasible, right? We're trying tobuild an organization for the
long term. And we had thathealthy tension all the time of
well, I know I need some bodies,but they're not the right
(43:04):
people. And, we think we'remaking the right decisions for
the long term but that tensionis something I feel and I deal
with every day.
Jonathan (43:13):
With COVID-19, can I turn that question around then
on to as a startup founder andmanaging not just like you and
your co-founder's schedules, butnow you've got 55 people and
want to double that, how haveyou been navigating the
challenges of whether it'sshelter-in-place or you are
(43:36):
already sort of managing Iunderstand two locations in San
Carlos, California, Seattle,Washington what's been
working? And what may havefailed in that you sort of
figured out like, ah, thatdidn't work to make it better
for running a company in thisenvironment?
Sujal Patel (43:51):
Yeah, I think that this has been an area where we
have spent a lot of time andenergy and we've been pretty
intentional. And you have tokind of look at our two offices
separately, because all of ourwet science is done down in the
Bay Area. And in Seattle, it'slargely software engineering and
(44:12):
administrative functions. I'm uphere, my head of finance up
here. And those are things thatcan be done from home pretty
effectively. So kind of if youlook at the Bay Area, once we
got to the shelter-in-placeorder in California, we did what
all companies had to do, whichis we shut down and we assessed:
what do we need to do as anorganization to build a safe
(44:35):
environment where lab workerscan work once it's okay
according to our county and ourstate regulations? And for us,
that meant that we had to spenda lot of time energy and money
on improving our office so thatit could be safe in this COVID
world. We added UV sterilizersin our HVAC systems, which was a
(44:58):
huge amount of work andexpensive. We added plexiglass
and glass partitions betweenevery office in our open
cubicle area, in our labs. Wespaced things out, which made
our space utilization much lesseffective. And we created
policies to ensure mask wear.
And we're created policies toensure hygiene, we created a
(45:21):
schedule for electrostaticsterilization for our
facilities. Now, one thing afterthe otherandthat meant we were
able to get back to somewhateffective work in three weeks,
and then get back to prettyeffective work within the first
two months. Even today, we still
deal with issues (45:44):
childcare, people are on weird schedules
because of their own childcareissues. We are very conservative
with respect to employees ifthey're exposed, or even if
they're two degrees ofseparation exposed to COVID. And
(46:04):
so if your roommate's boyfriendhad COVID, we don't want you
back at work until you've hadsufficient tests that say that
you're negative. And so we'vetaken this conservative
approach, and I think what it'sdone is it's helped us to
maintain some goodwill with ouremployees. You know, up in
Seattle, we were closed for alonger period of time, but
(46:28):
eventually a number of us wantedto return back to the work
office, and today we let peopleup in Seattle work from wherever
they would like (46:36):
from their house is fine, or from the
office is fine. And so you know,at any given day, today, there's
four or five, I think there'sfive of us here, but we have
quite a bit of space relative tofive people. And so we're able
to physically distant, we havesome UV sterilizers. And we kind
of are smaller groups that wereable to kind of keep our own
(46:58):
pot. So that's how we've stayedsafe. You know, what I would
say, we could have done betterand I hear this from other CEOs
as well as is that this pandemicis... it's a really scary thing
for people, it's unnerving, it'screating stress. And I think
that, we tried to move veryquickly. And we probably could
(47:18):
be even more intentional withour communications, with our
discussions one on one withemployees about their concerns,
and make sure that they are allkind of following our journey
one step at a time. And I thinkwe did a great job. And I think
that that's always an area wherewe can improve, right?
Jonathan (47:37):
That's... yeah, that's absolutely important, is the
communication and being present,and so that you can really have
your hands on the issues andtrying to lead the ship, as
well. What are your prioritiesthen... looking about a year
out? What do you prioritize overothers for Nautilus Bio?
Sujal Patel (48:02):
Yeah, so for us, there are, broadly, a few things
that we need to do, and they'revery interrelated. We need to
get to the point over the courseof the next 12 to 18 months,
that we have the ability toroutinely take customer samples
(48:23):
in and return valuableinformation to them, that they
have no way to analyze and getthat information anywhere else
in the world. And to startbuilding the body of evidence
that our technique is reliableand returning correct results
reliably. That's job number one.
In order to do that, I have todouble the size of my SAP
overnight. It's an interestingphenomena... a year ago, year
(48:47):
and a half ago, you'd run anexperiment, you'd get the
results, you'd figure out whatto do next. We're not in
research mode largely anymore,we're in engineering mode,
optimizing the protocol,figuring out how to build the
next version of the thing thatwe have, figuring out how to
make this thing go 50% faster tomeet the spec that it needs. And
(49:08):
in that world, there's 100experiments in every single
person's bucket that's piled upthat needs to get done. And so
we just need to grow ourcapacity, and we've got to be
able to build our engineeringand our research and development
organizations to be able to getto the goal of going on the
(49:28):
product side. On the commercialside of our business, we're
really focused on refining ourbusiness models, widening our
engagement with customers andsigning the initial deals with
customers that enable us tostart to engage in some
meaningful areas that willdemonstrate the capabilities of
our platform and those arethat's kind of the stuff that is
largely keeping me busy rightnow. When I look forward, one of
(49:52):
the big goals for me over thecourse of the last year has been
to transform the company from aleadership perspective from an
early stage developmentorganization to a company that's
ready to go and introduce aproduct to the commercial
markets. And so we entered 2020,with Parag and myself being the
(50:13):
only two executives at thecompany. And over the course of
the year so far, we hired MaryGodwin, who runs the operations
organization for us. She's beenat this for decades and has
worked with me across a numberof companies. We hired Nick
Nelson, who I talked aboutearlier. And then I've already
hired two other executives thatare starting between now and the
(50:35):
end of the year. And so we'regoing into 2021 with a team
that's really set up to help usto move Nautilus into a
commercialization phase.
Forrest Meyen (50:46):
I had a quick question, before we get the
tying things up... so I wantedto ask you what's been the most
fun part of your experiencebuilding this company?
Sujal Patel (50:58):
What's been the most fun part of building this
company? So, I, I am a computerscientist at heart. I am an
engineer at heart, the most funfor me, has been actually
getting involved and solvingreal problems and doing real
engineering work. This company,my experience is actually quite
different than my last company.
And so when I found that Isilonin January of 2001, I was 26
(51:22):
years old, and no single thingabout being a CEO. And I had to
grow my staff from me to whatended up as 35 people at the end
of year one because we had amassive chunk of software to
tackle, to go and do what we'retrying to do at Isilon. Because
I didn't know how to do anythingas CEO that consumed all of my
(51:42):
time and that's what I did. Atthis company, Parag and I
started with just the twofounders for the first six or
seven months, working throughthe tough challenge of hey, is
that algorithm gonna work? And Itook on a very different role,
right? I mean, I've been a CEOfrom miniscule through public
(52:02):
company, that stuff took 10% ofmy time, it's super easy for me
to do. I spent 90% of my time onthe biology and chemistry,
education, and on coding, andactually working through the
algorithmic pieces. So Parag wasin the wet lab, six, seven
months, when it was just the twoof us, and I wasn't, I was in
front of a computer coding. Andit was a ton of fun. And believe
(52:25):
it or not, even two to two and ahalf years in the company's
life, I held on to doing some ofthe coding myself, which was a
refreshing change that I did notget to do at the last place.
Forrest Meyen (52:37):
So the coding skills didn't just fade away as
you're many years as CEO at theother company?
Sujal Patel (52:43):
They didn't fade away. I mean, I've done plenty
of side projects along the way,even as a CEO, but things those
things come back to you veryquickly. But now, I don't do any
more coding, because there's ahuge team of software engineers
here now. But I still getinvolved in a lot of the
engineering aspects of whatwe're building and how we're
(53:03):
going to move that forward. Andit's a really fun part of what I
do, right? I mean, there's a lotof things that are part of a
CEO's life that are mundane, orstressful or hard, or
people-oriented, or whateverthey might be. But for me, the
engineering stuff is reallyenergizing and fun.
Jonathan (53:23):
To conclude the episode, we'll be glad to give
you a space to share anythoughts or comments or
shameless plugs that you'd likeour audience to hear?
Sujal Patel (53:36):
Shameless plugs?
Well, look, I think I think thatI've probably given you enough
shameless plugs here. Hopefully,what I've been able to convey is
that I think this is a hugelytransformational company in the
making. And I am incrediblyexcited and passionate about it.
And I really appreciate you,Jonathan, Forrest, giving me the
(53:59):
opportunity to go and talk toyour viewers and tell you a
little bit more about what we'redoing. And I encourage you all
to stay tuned, because I thinkthat over the course of the next
year or two, this company isgoing to be announcing some
really interesting stuff that'sgoing to have real impact on
human health and humanhappiness. And that's, that's
why I'm in this.
Jonathan (54:20):
Oh, that's great to hear. So exciting and Nautilus
Biotechnology for the audienceis hiring. So if it looks like
you've got the skill set andchops, and working with Sujal
and his team, then it could be agreat home for you.
Sujal Patel (54:34):
I'm Sujal Patel, founder and CEO of Nautilus
Biotechnology. Stay tough!
Jonathan (54:39):
And that's a wrap of probing the proteome with Sujal
Patel of Nautilus Bio. If youare a loved one suffer from
allergies, our next episode maybe right up your alley. We sit
down with Conor Cullinane ofPirouette Medical. He and his
team are developing anautoinjector drug delivery
system. Their first offering isan epinephrine auto injector for
(55:00):
emergency treatment of severeallergic reactions. Subscribe
and join the mailing list so youstay in the loop. Meanwhile,
stay tough.
Popular Podcasts
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Therapy Gecko
An unlicensed lizard psychologist travels the universe talking to strangers about absolutely nothing. TO CALL THE GECKO: follow me on https://www.twitch.tv/lyleforever to get a notification for when I am taking calls. I am usually live Mondays, Wednesdays, and Fridays but lately a lot of other times too. I am a gecko.