The new paper submission: “Resource ratio fluctuations drive the evolution of microbial metabolic strategies” (Zihan Wang, Jacopo Grilli, Akshit Goyal, Sergei Maslov) https://www.biorxiv.org/content/10.64898/2025.12.18.695187. In this project, we look at how fluctuations in the ratios of resources can steer evolution between specialist and co-utilizing strategies, with lag times playing an outsize role in what wins.
The earlier preprint https://www.biorxiv.org/content/10.1101/2024.06.14.599039 is now blessed with a fresh set of reviewer opinions from PLoS Comp Bio. Thankfully, they all seem pretty straightforward to address. Zihan and I will start working on it in January.
Submitted: another paper applying noSpliceVelo to flu-infected human cells
Another successful application of noSpliceVelo to explain gene trajectories in flu-infected human cells generated by Chris Brooke’s lab. The paper is submitted to a journal, while the biorxiv preprint version is here.
This figure show several endpoints of cell trajectories according to noSpliceVelo with one of them (terminal state 1) being highly correlated with high expression of the IFNL1 gene.
Submitted: ML-assisted GWAS in fatty liver disease
Fingers crossed! Kaushik and Ananthan just submitted our paper on ML-assisted GWAS in fatty liver disease (MASLD). The similarity of this disease’s abbreviation to Sergei’s last name is purely coincidental :-). It is a multi-journal submission to American Journal of Human Genetics (IF 8.1), Human Genetics and Genomics Advances (IF 3.6) and Patterns (IF 7.4). This work was done in collaboration with Sharon Donovan (PNI, UIUC) and several colleagues from the Mayo Clinic (Arjun Athreya, Alina Allen, and Konstantinos Lazaridis). BioRxiv and MedRxiv preprints are to follow soon.
Our auxotroph paper has finally been accepted in Cell Systems!
Our auxotroph paper with Tong Wang and Ashish Bino George has finally been accepted in Cell Systems! It went through a long review process, but better late than never. The preprint is available here.
Our paper on flu-infected human lung cells has been accepted in PNAS!
Our paper with Tarun Mahajan, Joel Rivera and Chris Brooke on flu-infected human lung cells has been accepted in PNAS! This is the first application of our own noSpliceVelo algorithm which is in its second round of review at Nature Genetics. The preprint is available here.
Interview with Sergei at NITMB website
Sergei gave an interview to s National Institute for Theory and Mathematics in Biology. Looking forward to visiting this place in the next year: their postdoctoral associates (and the views from the building) are spectacular. Check out the recordings of their seminars and conferences. I listened to a recent talk by Rama Ranganathan and was thinking about its connections to our epistasis paper.
We’re In! $33M ARPA-H Grant to Work on Phage–Bacterial Interactions
Big news!
We just got funded as part of the $33M ARPA-H MIGHTY project on phage–bacterial interactions!
Our group (together with Olgica Milenkovic’s lab) will focus on using machine learning to predict which phages go after which bacteria—and modeling of how phage cocktails evolve and interact over time.
Group Meetings Resume on Mondays 1:30-4:30pm Ct!
Our weekly group meetings are back for the fall semester on Mondays, 1:30–4:30pm CT.
Here’s the suggested order of presentations:
– Zihan (9/29)
– Xiaocheng (10/6)
– Kaushik (10/13) rehearsal of his BIOE qual presentation
– Ankit (10/20)
– Owen (11/3)
– Shreya (12/1)
– Kaushik again (12/8) This time a real paper
Who else wants to present? (12/15)
Sergei is giving two talks this week
Shameless Self-Promotion Alert:
Come watch me attempt to explain microbial crossfeeding and auxotrophy to mathematicians at the UIUC Math Bio seminar—Wednesday (10/1), 1pm CT, English Building 108, UIUC.
I’ll also be giving a talk at the NITMB Seminar in Chicago—Friday, October 3rd, 10:00–11:00am, 875 N Michigan Ave, Suite 3500. But let’s be honest: Chicago is a long way to go just to see me talk about microbes. Stop by my office any day if you want the same show—no travel, no line, and it’s always free!
Submitted: Protein Language Models Capture Structural and Functional Epistasis in a Zero-Shot Setting
Can protein language model predict protein fitness and spot protein epistasis without training on a new dataset? We put them to the test with real protein mutations—see what happens when deep learning meets experimental biology.
Lab members: Ananthan Nambiar, Sayantani B. Littlefield, Carlos Cuellar, Rohit Khorana, Sergei Maslov
Read the paper