Anna G. Green

Excellent work by Jatin Nainai (nainanijatinz.github.io), Bryn Reimer (code.brynmarie.com), and Connor Watts (connorwatts.github.io)
in collaboration with David Jensen (groups.cs.umass.edu/jensen/)

One circuit discovered in our protein language model

This work builds on recent papers that have used SAEs to uncover interpretable concepts learned by pLMs. We use activation patching to ask: which concepts are causally necessary for the model to perform contact prediction? For our two case studies, we find an explainable, sparse circuit

Figure 1 from Nainani et al, bioRxiv: activation patching for protein language models.

New work from my group using the tools of mechanistic interpretability to dissect the contact prediction capabilities of protein language models. www.biorxiv.org/content/10.1...

The premier conference on Machine Learning for Computational Biology is Sep 9-10 at the NY Genome Center in NYC!

Submission deadline is June 1 for 2-page abstracts and 8-page papers (eligible for proceedings track).

Registration is now open! (Link below)

Please retweet!

Excited to get started!

This was a huge collaborative effort with my talented PhD student @nvlsen.bsky.social, Irene Lepori, Zichen Liu, Shasha Feng, and the labs of @siegristpalmore.bsky.social, Marcos Pires, Joel Freundlich, and Wompil Im.

Identification of chemical features that influence mycomembrane permeation and antitubercular activity Tuberculosis (TB), caused by Mycobacterium tuberculosis (Mtb), is the deadliest single-agent infection worldwide. Current antibiotic treatment for TB takes a minimum of four months, underscoring the n...

Ever wonder why TB is hard to treat with antibiotics? One reason might be its unique outer membrane, which is a barrier for organic molecules. In our latest study, we used a combination of experiments, ML, and cheminformatics to understand membrane permeability in MTb. Preprint: shorturl.at/ALxvO

Antimicrobial resistance: a concise update Antimicrobial resistance (AMR) is a serious threat to global public health, with approximately 5 million deaths associated with bacterial AMR in 2019. Tackling AMR requires a multifaceted and cohesive...

The overview of AMR that you always wanted (someone else to write) - amazing!

Truly brilliant paper with great figures! @lancetmicrobe.bsky.social

www.thelancet.com/journals/lan...

Excited for this meeting!

Genomic Foundationless Models: Pretraining Does Not Promise Performance The success of Large Language Models has inspired the development of Genomic Foundation Models (GFMs) through similar pretraining techniques. However, the relationship between pretraining performance ...

Genomic Foundationless Models: Pretraining Does Not Promise Performance

I've long believed genomic foundation models are not as useful as claimed. In my mind, there isn't enough training data to justify their size. Interesting to see more work in this direction.

www.biorxiv.org/content/10.1...

White House pauses all federal grants, sparking confusion The Trump administration has put a hold on all federal financial grants and loans, affecting tens of billions of dollars in payments.

I think some people hear “grants” and think that without them, scientists and government workers just have less stuff to play with at work. But grants fund salaries for students, academics, researchers, and people who work in all areas of public service.

“Pausing” grants means people don’t eat.

My thesis work on active machine learning to model regulatory DNA is now out in Cell Systems!

We answer the question: When you can synthesize any DNA sequence you want, how do you decide which ones are worth testing?

www.sciencedirect.com/science/arti...

Antimicrobial resistance–attributable mortality: a patient-level analysis AbstractBackground. The impact of antimicrobial resistance (AMR) on death at the patient level is challenging to estimate. We aimed to characterize AMR-att

#AMR is a risk to antibiotic failure & death, but if this happens, do we record it on death certificates? Our centre data = NO!

👉In 1 year, 4% of deaths were AMR-attributed & NONE were recorded on death certificates!👈

Need to quantify this better to increase awareness! #IDSky @jac-amr.bsky.social

Can LLM agents discover novel protein functions? Introducing Gaia Agent 🌎 🤖: an AI biologist capable of reasoning across genomic contexts to predict functions of proteins! Gaia Agent is now integrated with Gaia Search at gaia.tatta.bio

Thank you Zam! Cool to see the paper featured. I agree with the writer that attention blocks and changes in padding strategies could be useful

Just announced: 2025 Computer-Aided Drug Design (CADD) GRC program

The topic: "Exploring the Synergy of Machine Learning and Physics-Based Computational Chemistry to Accelerate Drug Discovery"

It is shaping up to be a seminal conference. Hope to see you there!
www.grc.org/computer-aid...

Nice to see this systematically evaluated. Particularly interesting is the variable performance across different measurement types

Here is a #compbio starter kit! go.bsky.app/QVPoZXp To all the #Bioinformatics #Genomics #MachineLearning folks: please RP and let’s build this together!

When I started researching TB, one of my family members asked, "you mean the disease from the 1800's?" TB remains a huge public health crisis that deserves more attention.

Incredible graphic - the myriad of ways that bacteria defend themselves against antibiotics.

14 resistance mechanisms, summarised by Idan Yelin & Roy Kishony in Cell

www.sciencedirect.com/science/arti...

I think Pack 2 is up to 135/150 now, so between the two you can follow 285 bioml researchers in about 4 clicks!

Have since learned about this one: bsky.app/starter-pack...

I've made a starter pack for tuberculosis/mycobacteria researchers: go.bsky.app/Qvop3pP

Please DM me to get added or nominate others!