1/ In an excellent Nature Methods editorial, Lin Tang outlines the future of AI agents in biology. www.nature.com/articles/s41...

When you train your AI on genomes (ie. genome language models), what is the best split for train/validation/test? Most people split chromosomes, but is that the best way? 1/3

I’m recruiting a postdoc to work on algorithms for cancer genome reconstruction. We have access to a rich set of tumour samples sequenced across multiple technologies. If interested, feel free to DM. Please share.

On my way to @neuripsconf.bsky.social in San Diego ✈️ Really excited that Joel Bonnie and Tin Vlasic will present on Dec 7 at the ML4LS workshop: “From Base Pairs to Functions: Rich RNA Representations via Multimodal Language Modeling.” 🔗 neurips.cc/virtual/2025... 1/4

AI-for-biology is becoming one of the toughest places to publish. Output in mainstream ML is exploding, but biology-focused AI moves slower—and students feel the pressure first. 1/12

🚀🧬 Beyond Structure-based Biomolecule Design Its an important moment for structure-based biomolecule design: models starting to work and action shifting from academia to industry. So what are the next scientific problems academia could be thinking about? chaitjo.substack.com/p/beyond-str...

Well deserved. Richard is a giant in the fields of bioinformatics and computational biology.

Saturday mornings are ideal for contemplating AI in biology. It’s one of the most important frontiers for AI—but we can’t just copy what works for text or images. Biology speaks a different language. Right now, two major directions are emerging... 1/8

#GI2025 Mile Sikic @msikic.bsky.social presents "Geometric deep learning framework for de novo genome assembly" Now published in GenomeResearch @genomeresearch.bsky.social Full text here ➡️ tinyurl.com/Genome-Res-2...

Mile Sikic @msikic.bsky.social presents "AI for genomes—Rethinking de novo assembly" genome.cshlp.org/content/35/4/839 They devised a bidirectional message-passing procedure in GNN for the problem of genome assembly

Singletrack: An Algorithm for Improving Memory Consumption and Performance of Gap-Affine Sequence Alignment https://www.biorxiv.org/content/10.1101/2025.10.31.685625v1

Cool paper new paper from Lorién López-Villellas, @santiagomarco.bsky.social and others! Super cute and simple idea: In Gotoh's affine-cost alignment, only the M matrix is needed during tracing: we can just search for a gap-length x such that M[i][j] = M[i-x][j]+o+x*e or M[i][j] = M[i][j-x]+o+x*e.

🚀 Looking for talented PhD students! Join us in 🇸🇬 Singapore for 1-2 years to push the frontiers of AI for Genomics. Work on: 🧬 Cancer genome reconstruction 🧫 Cancer genome & cell foundation models 💊 RNA drug & mRNA therapeutic design #AI #Genomics #PhD 1/5

A popular humorous quote in the bioinfo community "About ten years ago, some computer scientists came by and said they heard we have some really cool problems. They showed that the problems are NP-complete and went away!” Now is the time for ML experts. 1/2

🚨New preprint out! 🧬Short reads can now decode centromeres. 🌍We reveal population-scale centromere haplogroups and their links to disease. biorxiv.org/content/10.1... (1/n)

🚀 AI + Genomics in Singapore! The AI Visiting Professorship at GIS invites world-class AI researchers (academia/industry, outside SG) to spend 20% of their time on AI–genomics projects. 💰 Up to 4M SGD (~3.1M USD) over 3 yrs. 📩 DM me if interested! #AI #Genomics #Research

Happy to present the first complete genome of an Indian individual biorxiv.org/content/10.1.... Joint effort by @astar-gis.bsky.social, FER, and partners, with support from the Singapore NPM programme! Joint effort with JJ Liu Lab! Prasad Sarashetti and @jlipovac.bsky.social as co-first authors.

Our RiNALMo RNA language model has been published in Nat Comm nature.com/articles/s41... Great work by Rafael Josip Penic and Tin Vlasic, with support and patience in teaching us RNA biology from Roland G Huber and Wan Yue RiNALMo is already an SOTA and a benchmark for RNA LLMs.

Seeking potential PhD students - Developing novel RNA therapeutics using AI-driven approaches - Decoding cancer genomics using LLMs and GNNs Deadline 1/6/2025. Contact me directly! a-star.edu.sg/Scholarships...

Work with @msikic.bsky.social, @rvicedomini.bsky.social, Kresimir Krizanovic MADRe is open-source, modular, and ready to use. Check it out: 🔗 github.com/lbcb-sci/MADRe 9/9

Join us at our GIS Seminar by Dr Marcin Magnus, Harvard University on OpenRNAFold: an end-to-end approach for RNA 3D structure prediction. Date: 29 April, 2PM Venue: GIS, Level 2 Host: Ivona MARTINOVIC @msikic.bsky.social @lvrcek.bsky.social

Join us at Genome Institute of Singapore! Great environment, stable funding!!

Apply now and be part of a team that values innovation, collaboration, and excellence! Check out www.linkedin.com/jobs/view/42... for more details. Find out more about the exciting work we do by our team – @jonathangoeke.bsky.social, @msikic.bsky.social

The first Singapore Long-Reads Symposium was a success! A massive thanks to the speakers: Keynote Heng Li, Ken Sung @jonathangoeke.bsky.social , @msikic.bsky.social & industrial partners @nanoporetech.com , @pacbio.bsky.social , @awscloud.bsky.social MGI_Technology Next Level Genomics & NovogeneAIT

GENOME: THE GIS SPEAKER SERIES Join us at our next seminar by Dr. Heng Li, Harvard Medical School & Dana-Farber Cancer Institute on Wed 12 March, 9.30am (SGT, GMT +8). Host: @msikic.bsky.social Webinar ID: 985 9424 0029 Participant ID: 371683 Passcode: 993251 #GISSPEAKERSERIES #genomics

We are organising ai4na-workshop.github.io at#ICLR2025. Topics: RNA, DNA and cell LLMs, structure, modifications, correction, variant calling... Deadline: Feb 10th 2025 Time for revision: 1 month Accepted papers - the opportunity to be invited by Nature Methods for submission

🚨Exciting news! We're partnering with @naturemethods.bsky.social 🧬High-quality submissions to the AI4NA workshop on @iclr-conf.bsky.social can also be considered for publication in the journal! 🌟 👉More details on our web page (link in comments) #AI4NA #ICLR2025

EXTREMELY cool and demonstrates clearly the promise of deep learning + bio. www.nature.com/articles/s41...

🚨 1 month to go! 🚨 The submission deadline for the AI4NA workshop at @iclr-conf.bsky.social is fast approaching! 🧬 ✨ Submissions on OpenReview will open soon—stay tuned! ✨ 🔗 Learn more on our web page (link below 👇) #AI4NA #ICLR2025

Verkko2 paper is finally on the biorxiv www.biorxiv.org/content/10.1...

Evaluating data requirements for high-quality haplotype-resolved genomes for creating robust pangenome references. #GenomeAssembly #HaplotypeResolved #Pangenomes #Genomics #GenomeBiology 🧬 🖥️ genomebiology.biomedcentral.com/articles/10....

Our hitchhiking paper is out at genomebiology.biomedcentral.com/articles/10.... great collaborative effort with Prasad Sarashetti, @jlipovac.bsky.social, Filip Tomas and JJ Liu. Briefly, the nanopore is enough for high-quality human genomes. Recently, HERRO showed that even UL is sufficient. 1/3

A comparative approach that predicts any modifications in RNA (or DNA) using raw @nanopore signals. Although simple, It outperforms the current SOTA. academic.oup.com/nar/advance-... w/ IvanVujaklija, Siniša Biđin, Marin Volarić, Sara Bakic, Zhe Li, Jianjun Liu and Roger Foo

RNA structure prediction is still an open problem. Look at our benchmark results (including Alphafold 3!) biorxiv.org/content/10.1... w/ Ivona Martinovic, Tin Vlasic, Yang Li, Bryan Hooi and Zhang Yang

For phased diploid genome T2T reconstruction, HERRO-corrected UL nanopore reads github.com/lbcb-sci/herro with cross-linking information using Hi-C/Pore-C is enough. Is this enough for heterogeneous cancer genomes? Or do we need to increase the accuracy of reads even further? Something else?