31 December is here, the last time we’ll see the numbers 2024 on our calendar. So, I would like to share the major advancements and news made this year in computational biology and probably even some of my personal learnings.
Alphafold3 Released
Of course, the highlight of this year is the open-source-ish release of Alphafold 3 this year, the next generation of all atom biomolecular structure prediction.
- The main difference being the inclusion of diffusion module
Announcing AlphaFold 3: our state-of-the-art AI model for predicting the structure and interactions of all life’s molecules. 🧬
— Google DeepMind (@GoogleDeepMind) May 8, 2024
Here’s how we built it with @IsomorphicLabs and what it means for biology. 🧵 https://t.co/K7uxMxdNr8 pic.twitter.com/MJfJ6NCEFW
Many other open source versions too:
Protein Design Competition
- Largest In-silico protein design competition conducted by Adaptyv
- De-novo design using the current best protein design method BindCraft by Martin Pacesa stands out!
PROTEIN DESIGN COMPETITION ROUND 2
— Adaptyv Bio (@adaptyvbio) December 7, 2024
The long awaited results are here -- and they don’t disappoint!
📈 We doubled the number of proteins we tested in our lab from 200 to 400!
🧬 Out of those 400 proteins, 378 expressed (95% expression rate!)
🚀 Out of those 378 expressed… pic.twitter.com/8Cr9xknEqu
Molecular Dynamics World Record
- Fastest Molecular Dynamics simulation at 1.1 Million timesteps/sec using Cerebras WSE3
- Read the report here, it’s pretty unique.
⚡ 🏆 Two world records in the same day? Sure, why not!?
— Cerebras (@CerebrasSystems) November 19, 2024
At 1.1M timesteps/sec researchers from @SandiaLabs, @Livermore_Lab, and @LosAlamosNatLab and Cerebras set a NEW world record using a single CS-2 for Molecular Dynamics simulations – this is 748x faster than GPU-based… pic.twitter.com/v4byNHbcjT
Best Evolutionary Protein Language Model -> ESM-C
Introducing ESM Cambrian.
— Alex Rives (@alexrives) December 4, 2024
Unsupervised learning can invert biology at scale to reveal the hidden structure of the natural world.
We’ve scaled up compute and data to train a new generation of protein language models. ESM C defines a new state of the art for protein… pic.twitter.com/GS1ShMvadv
My 2024
- Got my 4th and 5th semesters done
- Been building my portfolio and this blog site
- Basics of web dev along the way
- Learned a lot about protein design and engineering
- A few mountain treks
- Properly set up my social media sites
- Few GitHub builds and contributions
- Trying out exercise and yoga
- Lots of new friends
- Reducing my entertainment screen time a lot
- Regular Twitter user
- Using lots and lots of GPU’s for everything
- More family time
BYE 2024! See y’all next year ✌