One article genre that I’ve enjoyed recently are listicles of ‘most promising’ companies. With the necessary caveats, I still think its a great way to get up to date with what is exciting to a field at any point in time without the bias that comes with a post-mortem. Arguably the variable that separates experienced and unexperienced investors, scientists, operators, etc, is time. Knowing everything that has happened in a field up until a certain point is a tremendous advantage because it provides reasons for why certain things work or don’t work. It allows you to understand whether something is truly novel and a step improvement or just wrapped up in better marketing. It teaches humility, allowing you to understand all the reasons why something can go wrong. Understanding successes and failures could be useful for landscaping and spotting deficiencies in a fields that need to be explored.
How does one become better? One of the best ways that people improve at chess is by reviewing past grandmaster games and understanding what they did right and wrong. Similarly, looking back at the past through case studies should be a major source of potential for improving your understanding of an industry. As an exercise for myself, I looked comprehensively at all Fierce 15 Biotech companies (all pre-IPO) that have ever been published (since 2003). There are 289 companies (some duplicates or removed for other reasons). I don’t think Fierce is in any way perfect at choosing the top 15 companies, but they give a flavor of what was interesting to the biotech/biopharma community at the given time. Here is an example article with post-mortem included. They also have lists for MedTech and Healthcare which I look forward to reading. Endpoints also has their Endpoints 11 series which is essentially the same as the Fierce 15 biotech series.
My goal in doing this is to understand what has worked and what hasn’t. I want to have working knowledge of what companies have been built and how successful certain science has been so that the next time someone asks me to look at a neoantigen cancer vaccine company, I can point to 5 that have failed in trials. Or, when Alexey Borisy launches a new company with a16z and $122 in a Series A for combination therapies, I can see that he tried doing the same thing in 2000. Each company is an experiment, and people aren’t stupid. The ideas we think are great and the experiments we’d like to try now have probably already been tried before in some sense so having historical knowledge of failures is an easy way to screen and avoid unnecessary work but also to zero in on exactly what is the new innovation that is being paid for.
I’ve compiled the data here in this Google Sheet. It includes the company name, year founded, year on Fierce 15, whether the company can be classified as a platform or not, description, traction at the time of inclusion on Fierce 15, outcome notes at current time (October 2022), Acquirer (if applicable), Current Valuation, and Drugs Produced. Most of these are manually annotated so there may be certain biases or missing information/context.
Out of 289 companies, 177 (61%) are product focused, while 112 (39%) are platforms. Surprisingly, this hasn’t changed too much in recent years.
58 companies have not had anything to show for their work before shutting down, 163 have yet to produce a product (though still operational), and 68 companies have one or more products. These include diagnostics (FoundationOne CDx and Liquid CDx, OncotypeDx), sequencing platforms (PacBio HiFi sequencers), and 75 marketed drugs.
On average, it takes 4.06 years from founding to getting included on a Fierce 15 list, and a similar amount of time to exit after being included on the list (4.16 years).
Of companies that were acquired, the average acquisition price was $1.46 billion, though this is skewed by larger exits like MyoKardia ($13.1 B), Acceleron ($11.5 B), and Juno ($9 B) among other large buyouts. 32% of these are platforms, while 68% of these are product plays.
There is maybe more interesting data analysis to be done here, but it is less informative than case studies. There is an impossible amount of complexity to decode.
I tried to categorize companies into archetypes but this is a losing game. Each one is unique and there is so much overlap across categories that any resulting analysis would be pointless. The lesson here is that playbooks for certain archetypes are hard to come by. It’s difficult to say one model works better than any other. The beauty of capital markets is that each opportunity gets evaluated on an individual basis, and investors get to decide how much to pay and adjust accordingly if things are or aren’t working. The three archetypes I thought were relevant were asset centric rollups that are usually done in an indication specific manner based on a specific therapeutic vulnerability, exploration biology companies usually founded to explore targets in a new field of science, and new technology platforms to go after old targets in new more powerful ways.
However, a cool unanticipated thing that comes out of looking through so many years of data is that you can see clearly what ideas have gone in and out of fashion, as well as the evolution of certain ideas and the change that has occurred due to technological advancements. These lists have enabled longitudinal tracking of scientific revolutions.
There were 4 oncolytic virus companies starting with BioVex in 2005 which successfully developed Imlygic, the melanoma therapy approved by the FDA in 2015. The subsequent companies Oncorus (2016), PsiOxus (2017), and Turnstone (2017) have all yet to produce a drug in late stage trials. While these companies have somewhat fallen out of favor the past 5 years, there have been signs of life. CG Oncology has posted promising data to treat bladder cancer.
Cancer vaccines were another frequent area which included Immuno-Designed Molecules (2003), Therion Biologics (2004), Aduro Biotech (2014), Neon Therapeutics (2016), and Gritstone Oncology (2017). We’re in wait and see mode with Neon (acq. BioNTech) and Gritstone, but the earlier companies are clear zeroes. Recently, attempts at cancer vaccines are platform driven, less focused on antigen discovery. Moderna and BioNTech think mRNA is the answer, and other earlier stage companies like Attivare take a more complicated biomaterials based approach.
Another area that has fallen off excitement wise has been stem cells. Cellular Dynamics (2008) developed a screening platform using stem cells that was bought by FujiFilm in 2015 for $307 M. iZumi Bio (2009) was also engineering stem cells for disease modeling and therapeutics but it didn’t seem to work and they merged with Pierian. Finally, Cellerant Therapeutics (2005) was purifying HSCs as a therapeutic. This went to zero. Hype in the stem cell field has deservedly dried up, but new upstarts like Cellino and established companies like Fate may be able to shift sentiment.
One area that commands lots of attention these days is precision medicine. The common trope is that precision medicine has only been tried in oncology and only if we took a large scale precision approach to other disease areas cardiology, neurology, immunology, etc. we could address them piecewise. This is the thesis behind companies like Endpoint Health and Neumora. Blueprint Medicines (2011) and Myokardia (2015) have been the two big exits thus far. Blueprint is one of many, and while Myokardia brands itself on being a precision medicine company, the reason for its $\13 B buyout was for mavacamten which is not indicated for a specific patient population so it is unclear if its success was due to precision style drug development. Other attempts such as ARCA Discovery (Cardiology, 2006, $30 M market cap), BlackThorn Therapeutics (Neurobehavioral, 2017, acq. by Neumora), Goldfinch (Renal, 2020, TBD), and QurAlis (ALS/FTD, 2020, TBD) are ongoing and may provide an indication of the utility of this approach past oncology. The reason why oncology works well for precision medicine is that it is an acute condition with heterogeneity and rapidly dividing cells that is biopsied frequently. Contrast this with systemic disorders of the immune system that may be driven by genetics, but more likely environmental factors, and that are harder to perturb due to network effects.
Interestingly, many Fierce companies are focused on infectious disease. This field has suffered from financial difficulties and missed expectations.
One of the coolest parts of this study was looking at the progression of genetic medicines. Starting from Perlegen (2003) which set out to sequence and find drug targets from 50 human genomes, the incredible growth in our ability to sequence faster and cheaper has translated to company creation leveraging the sequencing revolution. Genomic Health (2004) was the first on the list to use genetic to diagnose human tumors, coming up with Oncotype Dx and selling to Exact Sciences. This was built upon by Foundation Medicine, whose FoundationOne test has been a core driver of the companion diagnostics evolution in cancer care. PacBio (2009) highlights the growth of new sequencing platforms which have continually lowered costs, increased speed, and delivered more accurate and high quality reads. Blueprint Medicines (2011) and Ultragenyx (2013) among others highlight the use of a wealth of sequencing information to understand disease biology and develop new targeted therapies for genetically defined conditions. The following iteration was delivering genes as a therapeutics product, with Bluebird (2012) and UniQure (2013) leading the way, and not long after came the commercialization of CRISPR to edit and design new products based on our understanding of genetics in basic cell and developmental biology. Now, almost all companies have used our improved knowledge of genetics and ability to perform genome engineering to develop new therapies and diagnostics. It’s the clearest example of a scientific revolution that one can find.
On the flip side, it was depressing to notice the many companies focused on infectious diseases and the poor economic outcomes even with good science. Out of the 17 companies highlighted by Fierce since 2003, there have been 11 drugs successfully developed and approved to be used. Several others are in late stage studies from Cidara, Spero, Vir, Ansun, Brii, and Atea. However, the exits on these companies have been disproportionately small. Corus, Optimer, Cerexa, and Idenix had healthy exits but their drugs have had poor commercial performance. MerLion, Tetraphase, and Rib-X all are either bankrupt or on the brink even though they have collectively developed 7 drugs. The late stage valuation of Cidara and Spero are far below raised capital in the micro-cap range. There is perhaps potential of this changing with coronavirus driven sentiment change. Noticeably, Adagio (now Invivyd), Atea, and Vir have all reached healthy valuations due to work on Covid-19.
Teasing out insights from data can be a losers game without specific hypotheses to test but I can come up with three:
The big name science doesn’t always work. Longevity (Sirtris & Elixir) was a big flop. Many many other examples in this list. On aggregate, I think this list has a better batting average than the industry, but at some level it becomes quite clear that no one knows anything.
Browsing through the failures of old should be a core part of any brainstorming. There is a database of failed companies compiled here. I hope there is a database or repository of failed drug candidates to do head to head comparisons. Testing against placebo in a clinical trial or even putting a drug in a clinical trial is quite scary. Hopefully past learnings become a core aspect of improving new drugs.
Someone should create a list of unsolved problems in biology and for relevant problems, a post-mortem on why certain experiments or companies went wrong. Recently heard a talk from Jay Bradner and was struck by how intimate of knowledge you need to have in order to excel and have an edge in the industry. A big portion of this is knowing what the problems are and when to try applying the newest technologies towards those problems. I think the answer currently is just to read more and talk to more people, but having a centralized and continuously updating database would be a tremendous reference. I would guess that a lot of top scientists just have this floating in their head.