Thread by @jchodera

1/ I really enjoyed giving the annual State of the Lab Address this week---a tradition shamelessly borrowed from my PhD advisor Ken Dill. It's a chance to step back and reflect on the major goals of the laboratory and refocus for the future. Here are some highlights:

2/ The lab continually strives to improve *everything*: Our processes, our code, our wetlab protocols, our approach to mentorship, even the processes to improve our processes. Nothing is perfect, but we can strive to improve by continually looking for opportunities to do better.

3/ We regularly evaluate ourselves against what our principles. The most important part of *my* job is to make sure my trainees achieve their goals to end up in good, satisfying careers. Doing this requires we excel in four areas.

4/ I've been extraordinarily lucky to get to work with such a talented crew of scientists both here at @sloan_kettering and in my new Berlin satellite group at the @ChariteBerlin as part of my BIH Einstein Visiting Professorship: choderalab.org/members

@sloan_kettering @ChariteBerlin 5/ I'm tremendously excited to head to Berlin in a couple of weeks for my first official visit as part of the BIH Einstein Visiting Fellowship, where I share two group members with the extremely talented Andrea Volkamer at the @ChariteBerlin: physiologie-ccm.charite.de/en/research_at…

@sloan_kettering @ChariteBerlin 6/ Over the last few months, it's also been an absolute joy to work with the large number of talented scientists involved in the @openforcefield Initiative, spread among five different research groups, all doing science in the open! Check out their work at openforcefield.org

@sloan_kettering @ChariteBerlin @openforcefield 7/ This past year, we graduated the first PhD student from the lab, Patrick Grinaway, who was with the lab from almost the very start! He's now working at onai.com here in NYC, alongside some some incredibly talented people.

@sloan_kettering @ChariteBerlin @openforcefield 8/ We also saw @GSKGradSchool PhD student Steven Albanese graduate in under five years! Steven is now a member of the Drug Discovery Applications Group at @Schrodinger putting his skills to good use in helping others discover new therapies.

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger 9/ We use a lot of tools to keep all of these folks together. @SlackHQ has been incredibly useful for coordinating across research groups (or continents!), as has @Google GSuite. And I ❤️ @overleaf for making manuscript writing a joy, and @github for software development.

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github 10/ We're always trying to improve our processes for onboarding new people. With two new technicians joining, our awesome lab manager Erin Grundy has prepared an Onboarding Trello to streamline the process (h/t to Christina Johnson for this idea!).

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github 11/ Sadly for us---but awesome for science!---our lab manager Erin Grundy will soon be leaving the lab to start a PhD at George Washington University. Congratulations again!

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github 12/ Picking up on the excellent suggestion of @jenheemstra (those of you who aren't following her yet are missing out!), I proposed we implement a written Lab Wellness Policy to clarify how the lab prioritizes the mental and physical health of its members.

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra 13/ Luckily, @jenheemstra was kind enough to publicly share her own Lab Wellness Policy with the world. The lab unanimously voted to adopt such a policy, and thought this was a great starting point.

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra 14/ We have a few additional lab policies and reminders we'll integrate into the policy. Our vacation policy is mostly just "let JDC know your plans and make sure to put it on the group calendar so we don't bug you when you're away".

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra 15/ Our core mission is to do awesome science and communicate it to the world. We take time each State of the Lab to remind ourselves what our scientific objectives are, both in terms of goals and the way in which we do science.

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra 16/ One of the major scientific outputs of the lab is software tools that explore ideas of how physical modeling could aid drug discovery and basic research into biological mechanisms at the molecular level. We strive to have *impact* in enabling others to do more science.

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra 17/ Some of our code is much more experimental, but all of it is available on github.com/choderalab
Some of the more popular packages are indexed on our group webpage at choderalab.org/code

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra 18/ To maximize impact, we want to ensure our work can be used, modified, built on, and redistributed. We've been inspired by the work of @victoriastodden in describing the legal framework behind the "reproducible research product". Check out her paper: web.stanford.edu/~vcs/papers/Li…

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra @victoriastodden 19/ We're also very much inspired by the culture in machine learning that celebrates reproducible research, as evidenced by sites like paperswithcode.com; as @wpwalters says, "just put the code on github!" is a great way for others to try it, understand it, and build on it.

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra @victoriastodden @wpwalters 20/ We're also inspired by the incredible productivity that abstractions have provided the machine learning community, where modern software frameworks that utilize these abstractions make it incredibly easy to do new science and test new ideas. e.g. tensorflow.org/overview

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra @victoriastodden @wpwalters 21/ While we may not have the resources of @GoogleAI, we're working to build frameworks that leverage useful abstractions for molecular modeling. Using openmmtools from @andrrizzi, you can build a whole alchemical free energy code in just a few lines! openmmtools.readthedocs.io/en/0.18.1/gett…

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra @victoriastodden @wpwalters @GoogleAI @andrrizzi 22/ You can read more about all of the features openmmtools provides in @andrrizzi's excellent tutorial: openmmtools.readthedocs.io/en/0.18.1/gett…

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra @victoriastodden @wpwalters @GoogleAI @andrrizzi 23/ One of the more exciting directions we've been exploring is how we can combine alchemical free energy calculations, machine learning, and automated biophysical experiments to predict clinical cancer resistance mutations at scale.

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra @victoriastodden @wpwalters @GoogleAI @andrrizzi 24/ There are some deep questions to be addressed in understanding how to successfully apply machine learning to this and related problems in drug discovery.

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra @victoriastodden @wpwalters @GoogleAI @andrrizzi 25/ We're excited about the @openforcefield Consortium, where pharma partners who understand how open science and open source software can have transformative impact across the field are funding key research for the benefit of all.
openforcefield.org/consortium

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra @victoriastodden @wpwalters @GoogleAI @andrrizzi 26/ The @openforcefield effort is an opportunity for us to lead by example in doing science in the manner we'd love to see other large collaborations operate, with an eye toward maximizing impact at every stage.
openforcefield.org

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra @victoriastodden @wpwalters @GoogleAI @andrrizzi 27/ We have ambitious and demanding goals to achieve with @openforcefield, but we can transform the way that molecular mechanics force fields are built, evaluated, and used in the process.

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra @victoriastodden @wpwalters @GoogleAI @andrrizzi 28/ It's been wonderful to work with @dga_smith and the @MolSSI_NSF software scientists as part of the @openforcefield collaboration, and the new public QCArchive that will hold all the quantum chemical data we generate will be a fantastic resource: qcarchive.molssi.org

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra @victoriastodden @wpwalters @GoogleAI @andrrizzi @dga_smith @MolSSI_NSF 29/ You can dive in and get started playing with live @openforcefield torsion drive data right now by launching the examples!
browse: docs.qcarchive.molssi.org/en/latest/basi…
launch in @mybinderteam: mybinder.org/v2/gh/MolSSI/Q…

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra @victoriastodden @wpwalters @GoogleAI @andrrizzi @dga_smith @MolSSI_NSF @mybinderteam 30/ The @openforcefield effort isn't just about producing papers or force fields. We've tried to thoughtfully identify how to share all of the research products we generate in the process in a useful manner.

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra @victoriastodden @wpwalters @GoogleAI @andrrizzi @dga_smith @MolSSI_NSF @mybinderteam 31/ One of the coolest aspects of the project is that @cbm_student Josh Fass has been working on ways to use statistically motivated reversible-jump Monte Carlo to automatically determine the number of atom types in Bayesian parameter fitting schemes! Here's a GBSA model example:

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra @victoriastodden @wpwalters @GoogleAI @andrrizzi @dga_smith @MolSSI_NSF @mybinderteam @cbm_student 32/ The first few generations of @openforcefield force fields will be produced with the awesome ForceBalance tool from Lee-Ping Wang [github.com/leeping/forceb…], but we're already thinking about how to make fitting a force field as easy as training a machine learning model.

@sloan_kettering @ChariteBerlin @openforcefield @GSKGradSchool @Schrodinger @SlackHQ @Google @overleaf @github @jenheemstra @victoriastodden @wpwalters @GoogleAI @andrrizzi @dga_smith @MolSSI_NSF @mybinderteam @cbm_student (short intermission)