🚨#COVID19 PAPER ALERT🚨
With your help, F@h became the first #exascale computer, collected 100,000X more data than a normal paper, learned how the #Demogorgon evades our immune system, and discovered drug targets called ‘cryptic’ pockets (#FAHPockets)!
biorxiv.org/content/10.110…
With your help, F@h became the first #exascale computer, collected 100,000X more data than a normal paper, learned how the #Demogorgon evades our immune system, and discovered drug targets called ‘cryptic’ pockets (#FAHPockets)!
biorxiv.org/content/10.110…
With over a million people downloading F@h, we reached #exascale computing speed within weeks. This made us >5X faster than the IBM Summit supercomputer, the fastest traditional supercomputer in the world!
This newfound power enabled us to simulate every possible protein from the #SARSCoV2 virus in parallel. Together, we generated the largest dataset of protein dynamics in history. We’ve highlighted some of our exciting findings in this thread:
We learned how the Spike #Demogorgon hides from our immune system. We were able to characterize the opening of the Demogorgon’s “mouth”, revealing the site that binds human cells and drives infection. Targeting these structures of the mouth presents therapeutic opportunities!
With #exascale computing we simulated spike #demogorgon proteins from 3 #coronaviruses: #SARSCoV2, the 2003 SARS1, and NL63. We saw all three Spike proteins open and found a correlation between Spike protein opening and a virus’ ability to bind a human host!
Going after the #demogorgon’s mouth isn’t the only option. We found multiple cryptic pockets not yet observed by experiments. Each pocket presents opportunities for drug design! For example, we found a pocket in the Main Protease that is distant from the “active site” (spheres):
Our simulations are already joining forces with experiments! We are collaborating with labs @WUSTLmed and our partners @covid_moonshot (which targets the Main Protease) to test predictions made by the data you all generated.
@WUSTLmed @covid_moonshot We are still hunting for pockets (like one shown here in NSP10) – as we find more #FAHPockets, we will update the manuscript to reflect what we learn.
The paper is available now on @biorxivpreprint for free: biorxiv.org/content/10.110…
@biorxivpreprint This data was generated by people around the world, and so it belongs to everyone. We’re working with our partners @MolSSI_NSF to share it online and will make announcements as data gets posted to osf.io/fs2yv/ and covid.molssi.org
@biorxivpreprint @MolSSI_NSF This work is a major milestone in our efforts to tackle #COVID19. We want to thank all our partners, including: @Microsoft #AIforHealth,@Avast, @Oracle, @VMware, @NVIDIA, @LinusTech, @PCMR, @awscloud, @cisco, @amd, @PureStorage, @UKRI_News, @Seagate, @intel, @arm, and @a16z!
@biorxivpreprint @MolSSI_NSF @Microsoft @Avast @Oracle @VMware @nvidia @LinusTech @pcmr @awscloud @Cisco @AMD @PureStorage @UKRI_News @Seagate @intel @Arm @a16z Also thanks to @Github, @DataDog, @GIGABYTEUSA, @NSF, @NIH, @CERN, and @Dell
A big thank-you to our host universities @WUSTLmed, @sloan_kettering, and @TempleUniv for their continued support! Also, shout-out to @wustlDBBS for supporting graduate students in the Bowman lab for much of this work!
@WUSTLmed @sloan_kettering @TempleUniv @WUSTLdbbs Lastly, a big *thank you* to our community. We wouldn’t be able to publish historic science like this without your contributions, and every result here is because of each of you. So thank you, stay tuned for our next set of results, and keep on Folding!
