C2Bio

Curated weekly digest of notable AI/ML and biotech developments. I really liked the paper " Guiding generative models to uncover diverse and novel crystals via reinforcement learning" by  Hyunsoo Park & Aron Walsh, published in Nature Machine Intelligence. They leveraged RL to guide the generative model to generate novel crystals that differ from the training data distribution, which also showed more stable training behavior. Traditionally model generate new data by sampling in latents space close to the original training data, which limits novelty, and may even hinder the discovery of entirely new structures that are no longer subject to evolutional selection. Very nice work.  Although I do not work on crystal structures, I do feel their work is inspiring for bio-centered AI/ML applications. 🧬 Biotech ★ TranscriptFormer: A generative cell atlas across 1.5 billion years of evolution — A generative foundation model trained on 112 million cells across 12 species ach...

I am starting a series of weekly digest of notable AI/ML and biotech developments related to my own interests. While most of the contents here are published in the past week, I may also add papers that were published earlier but just got my attention. I hope you will find these weekly digests useful. 🧬 Biotech ★ Multivalent mRNA vaccine platform with compatible antigens conferred broad-spectrum protection against orthoebolaviruses' exposure — A multivalent mRNA vaccine platform delivering glycoproteins and nucleoproteins in a single LNP for broad-spectrum protection against EBOV, SUDV, and BDBV. (PNAS) ★ Biomolecular profiling for noninvasive health monitoring — A review highlighting how MS-based molecular discovery and wearable sensing serve as complementary approaches for continuous, noninvasive health monitoring. (Nature Biotechnology) 🤖 AI/ML ★ Evaluating Large Language Models in Scientific Discovery — A scenario-grounded benchmark evaluating LLMs across bi...

  FlashABB is a fast antibody structure prediction library from the Oxford Protein Informatics Group. (OXPIG) with a preprint here . It predicts paired VH/VL structures, scores, developability (FlashTAP), and produces structure-aware sequence embeddings (FlashABB-SSS) — all from a single "HEAVY|LIGHT" sequence string, with weights bundled in the repo. The library was written for CUDA. Getting it running on Apple Silicon required fixing five separate issues. I worked with Claude Code and got the following fixes:   --- Fix 1 — pip ≥ 26 editable install pip 26 dropped the legacy setup.py-only editable install fallback. A minimal pyproject.toml was all that was needed:   [build-system]   requires = ["setuptools>=64", "wheel"]   build-backend = "setuptools.build_meta"   ---   Fix 2 — Device auto-detection   Three places hardcoded device='cuda'. Replaced with a helper:   def _default_device():       if torch.cuda.is_available(): ...

Great to see that DeepSeek R1 is published as a research paper, which allows people to understand more about the engineering and design decisions made by their incredible team.  According to a comment by the editor of Nature, this paper was reviewed by 8 experts. (Most scientific journals require only 3 independent reviewers.) This reminds me of AlphaFold, a revolutionary protein model that can fold most proteins sequence into 3D structures with accuracy comparable to crystallography.  AF was published as a scientific paper followed by a Nobel prize. However, a lot of details are missing in the paper, especially training methods, which is crucial for fine tuning the original model (if it is open sourced). Open file mimics AF with full open source and well documented training methods, which allows scientists to understand the model in detail and appreciate the engineering marble in addition to being able to fine tune the model. Love to see more publications of mode details ...

A fundamental challenge in generative artificial intelligence involves sampling from carefully constructed high-dimensional latent spaces and utilizing these samples as inputs to decoder networks for generating novel entities. In the context of computational protein design, this process typically involves sampling regions within protein embedding spaces where specific biochemical properties are anticipated, such as enhanced binding affinity or improved developability characteristics in therapeutic antibodies. The sample-decode paradigm presents several significant technical challenges that must be addressed for effective protein generation. First, determining the optimal sampling distance from training datasets remains a critical consideration—sampling too close may limit diversity, while sampling too far may compromise biological relevance. Second, identifying which directions in the latent space merit more extensive exploration requires careful consideration of the underlying protein...

After tracking Nvidia’s RTX 50-series cards for a while, I still feel they are too expensive and likely not powerful enough to fully support my AI enthusiasm. As a result, I’ve decided to explore alternatives, particularly cloud solutions. Google Colab has turned out to be a very appealing option for now. It provides a Jupyter Notebook interface backed by CPUs, GPUs, and TPUs managed by Google. There is a free tier that offers limited GPU access (availability-based). During my test run this Sunday morning, I was able to access a GPU with 16 GB of memory. Paid users receive access to more powerful GPU and CPU resources. This makes Colab a viable alternative to running an on-premise AI lab—provided you’re comfortable with Google hosting your data and models. I haven’t checked AWS’s GPU offerings yet, but based on my prior EC2 experience, I feel Colab is much more convenient. Colab simplifies many aspects (from billing to system management), allowing you, as an AI scientist, to focus on m...