Discovering Nimbus.

Posted March 10th, 2011 in New business models, Portfolio news

Bill Gates has just backed one our new startups – Nimbus Discovery LLC – as part of an extension to the seed tranche. Here’s the press release.

It might come as a surprise to some, but Bill Gates has been a long-time biotech supporter: he was a founding investor in ICOS and was on that Board for 15 years, and importantly, he’s also one of the largest investors in Schrödinger, the world’s leading computation drug discovery company, and our founding partner with Nimbus.

So, with this financing, we’re launching Nimbus out of ‘stealth mode’.  Here’s the story.

We founded Nimbus Discovery in 2009 with Schrödinger and have incubated the company here at Atlas for the past couple years. It’s fair to say Nimbus is a rather unconventional biotech, possessing three distinctive features.

  1. Unique Drug Discovery Partnership with Schrödinger provides access to an unparalleled technology suite without the financial burden of having to build it organically. Beyond the sheer breadth of Schrodinger’s software offerings, the crown jewel from our perspective is their new technology for evaluating the energetics of specific water molecules in binding sites called WaterMap™.
    • Our bodies are 90+% water and yet most structure-based drug design models fail to integrate proper solvent (water) effects with regards to both entropy and enthalpy. WaterMap™ does this. We’ve already found it to be an incredibly powerful tool for identifying specific optimization strategies based on novel water-energy-driven Structure-Activity-Relationships. WaterMap™ is a incredibly well validated technology and has been applied (retrospectively) to about 45 different targets using ligands that have highly heterogenous structures. Not only does WaterMap™  accurately predict binding affinities, it explains SAR that would otherwise be inexplicable. In short, super cool science at the cutting edge of in silico drug discovery.
    • Our partnership with Schrödinger provides us with far more than just this software package though – we get a large number of dedicated computational chemists, access to thousands of processors via their cloud computing network, new unreleased software algorithms, and exclusivity around specific targets. Continuous, advanced access to the most cutting edge technology applied in a personalized way to our targets allows Nimbus to maintain its first mover advantage.
    • Moreover, Schrodinger continues to make a huge investment in its platform leveraging an army of ~60+ PhDs and the deep-pocketed support of Bill Gates and David Shaw. (As an aside, it’s probably the only biotech with two billionaires as its top two investors). It is no surprise that Schrodinger has led innovation in the field: the company currently has 50+ peer reviewed publications many of which are among the most heavily cited articles in the in silico modeling space.
  2. Ultra-lean “virtually integrated, globally distributed” R&D operating model to aim for exquisite capital efficiency. We’re really pushing the envelope on virtual drug discovery.  We have 10-15x more FTEs working for the company externally as inside the company.
    • The core team is two incredible individuals (Rosana Kapeller and Jonathan Montagu) who, in addition to being very smart seasoned biotech startup veterans, excel at integrating remote workstreams and collaborators. We’ve got several teams at offshore CRO partners doing biology, chemistry, crystallography, in vivo work, etc…. Not to mention a core set of KOL academic collaborators. We think we’ll be able to work on 3-4 programs with this setup.
    • It’s paying dividends already: on less than $2M spent, we’ve generated a selective, potent IRAK4 inhibitor (cancer, inflamm) and a set of lead scaffolds against other targets.
    • One of the key features of the operating model is the virtual integration of all these pieces, with in silico model refinement in the core of the ‘engine room’ so to speak. WaterMap and tools like it depend on constant structural model enhancement, which requires real time integration of project data into our models. Our remote virtual teams interact on almost a daily basis to integrate these new streams of information. This allows us to use these tools for more than just improving selectivity and potency – but also to more precisely know which part of molecules to optimize for PK/ADME concerns as well.
  3. Novel asset-centric corporate structure to promote liquidity and capital velocity.  Back in 2009 we spent alot of time figuring out how to leave the limitations of the traditional C-corp structure behind and adopt a more flexible LLC-holding company structure with target-specific subsidiaries as C-corps.  A few companies have recently announced they are doing this as well, like Adimab and Ablexis.   This structure does a couple very valuable things (beyond creating a job for accountants  to track the project financing).
    • First, it enables project-driven ‘clean’ transactions with Pharma, such that a Pharma can just acquire the target-specific subsidiary and own the IP/assets of a particular program if it so chooses.
    • Second, and importantly, it solves the classic drug discovery illiquidity problem (where it takes 7-10 years to get liquid via M&A or IPO); this LLC structure enables us as shareholders to cycle capital back to our investors in a tax-efficient manner on a per project basis. Furthermore, it creates this type of ‘deal optionality’ without foregoing any of the traditional benefits of C-corps.

At the end of the day these three elements are great value enablers, but ultimately it’s about the medicines we discover. To pick the targets we sought to generate drugs against, we took an orthogonal approach to traditional “biology” driven target selection. History has shown that generally in silico technologies in drug discovery are helpful tools for the vast majority of targets, but not game-changing.  With Nimbus, we wanted to let the technology identify targets where its new insight into SAR was potentially transformational rather than incremental – essentially, to find the rare 10% of targets where these technologies offered a compelling path to new, differentiated chemical matter against hot targets.  To accomplish this, we spent the first 12 months of the company screening several hundred ‘hot targets’ in the industry’ pipeline before picking the 10% or so to focus on with Schrödinger.

Our two lead programs today:

  • IRAK4. One of the most interesting immune-kinase targets in both B-cell cancers (like the IRAK4-dependent MyD88-driven Activated Diffuse B-cell lymphoma [link]) and inflammation. It’s traditionally been very challenging to get selectivity and cellular potency; we’ve managed to generate a series that addresses both of those and aim for a development candidate by end of 2011.
  • ACC or Acetyl-CoA Carboxylase. A critical enzyme in the metabolism of lipids and a top target for obesity as well as cancer metabolism. It’s been very hard to drug effectively. We’ve managed to get very interesting leads against a unique allosteric domain that should enable a differentiated profile.

Saving the best for last, it’s fair to say our team is exceptional. Rosana Kapeller is our CSO and was founder of Aileron Therapeutics after nearly a decade at Millenium. Jonathan Montagu is our CBO and was with Concert, J&J, Chiron and BCG. The broader team of folks at Schrodinger, like Ramy Farid (President of Schrodinger and founding Board Director of Nimbus), and our chemistry leadership (Ron Wester, Gerry Harriman, Donna Romero, John McCall) are also incredible.

With that rundown, we’re officially out of stealth mode and aiming to close on a Series A soon.  Exciting times.

This entry was posted in New business models, Portfolio news. Bookmark the permalink.
  • Congrats! This virtual biotech model sounds very interesting. Do you think this low-cost model will bridge the gap between the “death valley” of academic IP and commercial drug development?

  • Anonymous

    exMBB, thanks for the post. I certainly think so – translating academic discoveries by leveraging a distributed group of expert CROs is very doable. Thx for the post

  • Anonymous

    Very nice structure, making the most of the VC advantages while leveraging the experience of Schrödinger. $2M for a potent IRAK4 inhibitor (+ scaffolds) sounds like a good deal, I’m hoping for more news soon. Regarding the “asset-centric corporate structure” you mention, is there any literature or other links for the interested-but-inexperienced you could recommend? Many thanks in advance

  • Anonymous

    Nothing published on the structure, but here’s the gist: Nimbus LLC is the parent company, all new programs are in underlying C-corps. The mgt team is also in a C-corp. All operating expenses are in the C-corps. When Pharma does a deal for a program, they buy the C-corp and future milestones flow through to shareholders.

  • Anonymous

    Many thanks! Will endeavour to keep uptodate with developments…

  • very smart! A similar model was tried in the mid 90’s by a UK company called Oxford Molecular. Recognising that their comp. chem tools sold at relatively low unit price, the company pulled together an interdisciplinary team of (academic) scientist and computational chemistry to undertake drug discovery on ion channels. They partnered the project with a Japanese pharmaceutical company for a seven figure fee.

    I wish Nimbus huge success.

  • J Patrick Nelli

    Great post. Nimbus seems to be pushing forward with an efficient model. Do you know how many in silico biotechs have sprung up in the last decade? It would be interesting to look at their success versus more traditional biotech startups.

  • Anonymous

    There have been a bunch of startups claiming to have a better computer program for drug design (Locus, Acceryls, etc..). Most have been disasters at least from an investment perspective. Nimbus is unique in that we are partnering with the world leader, rather than internally developing the technology. We also, unlike others, have let the technology tell us the targets it can distinctively contribute insight to, rather than pushing the hottest target of the day into a technology that might not fit. At the end of the day, the in silico engine is important, but the virtual operating model and the corporate structure are also major contributors to the Nimbus proposition.

  • I wonder how how soon will David Baker will start his own computation drug devel company. Though I am sure he’s not short of money, and he’s truly an Einstein-type scientist…He has the hair to go with it too

  • Pingback: Blog: Discovering biotech startup Nimbus - Tech News- Tweets()