Ivan Sovic

HERRO has been published in @Nature !! nature.com/articles/s4158… This achievement is a result of the great work by @domstanojevic, with contributions from @DehuiLin, @sergeynurk, and @PaolaFlorezdeS We are entering the era of high-quality genome assemblies supported by AI. This is just the beginning.

Hiring a PhD student in AI × RNA 🧬 • Predict, generate and optimise RNA 3D structure 📍 Zagreb (EU hub, great quality of life) 🌍 Team across Zagreb & Singapore (visit SG possible) 🧠 CS/math/physics + strong AI (not just prompting) 🧪 Real wet-lab validation of models DM if interested. #news_95608" target="_blank" rel="nofollow noopener">fer.unizg.hr/novosti/natjec… under item 6! Please share.

🌟 We finally have a brand new lab webpage, with blog-style articles describing our work, and exciting *open positions* in culturomics and metagenomics! Please spread the word 🙏 mtms-lab.github.io csb5.github.io/open-positions

Are neural nets across modalities really converging to the same representation as they scale, as the Platonic Representation Hypothesis suggests? We show that common representational similarity metrics are confounded by network width & depth. We propose a permutation-based null calibration that fixes this. Result❓ • Global convergence largely disappears. • Local neighborhoods persist. We propose the alternative Aristotelian Representation Hypothesis: Neural networks, trained with different objectives on different data and modalities, are converging to shared local neighborhood relationships Very proud of @FabianGroger and @ShuoWen18 for this work! Paper: arxiv.org/abs/2602.14486 Webpage: brbiclab.epfl.ch/aristotelian Code: github.com/mlbio-epfl/ari…

Excited to share @iclr_conf papers from our team! 📌 LaMer: enables LLM-based agents to actively explore and adapt via environment feedback by @YulunJiang🔗arxiv.org/pdf/2512.16848 📌HeurekaBench: a benchmarking framework for evaluating AI co-scientists, grounded in real experimental workflows by @sibasmarak🔗 arxiv.org/pdf/2601.01678 📌CW-PO: confidence-weighted preference alignment that requires minimal human supervision by @myeongho_jeon 🔗openreview.net/pdf?id=ROioaZ4… See you in 🇧🇷!

Great opportunity!! Fran is an outstanding PI

Happy to present our work on RNA large language models at 🚀 AI meets RNA Biology – Mini-Symposium 📅 Wednesday, December 17 | 📍 Hybrid (@EPFL_AI_Center & Online) 📍 Register for the in-person event here: forms.office.com/pages/response… 💻 Join online via Zoom: #success" target="_blank" rel="nofollow noopener">epfl.zoom.us/j/62574766054#… The meeting is organised by @ISBSIB and @unibern, and hosted by @EPFL. Thank you, @RLuisier, for organising and inviting me

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, San Diego. ✈️ Proud that Joel Bonnie and @TinVlasic will present on Dec 7 at the ML4LS workshop neurips.cc/virtual/2025/l…: “From Base Pairs to Functions: Rich RNA Representations via Multimodal Language Modeling.” I’ll also be giving a keynote on Dec 7 at @workshopmlsb , mlsb.io (co-located with @NeurIPSConf), discussing the next wave of RNA structure prediction. Open to conversations on: • #BioML for RNA structure prediction + optimisation • Improving genome and cell foundation models • AI-enabled genome assembly • Joining us at @astar_gis — we’re hiring PhDs, postdocs, engineers, and interns in AI just send me a DM

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... 1/6

We are seeking PhD students who would like to work with us on challenging problems in AI for genomics for one or two years. Problems include cancer genome reconstruction and the cancer genome and cell foundation models, as well as the optimisation of mRNA therapeutics and the targeting of RNA with drugs. Are you pursuing a doctoral degree at a university and looking for a short-term research attachment in Singapore? We are open to applicants for the A*STAR Research Attachment Programme (ARAP) - a-star.edu.sg/Scholarships/f… 🔍 About ARAP The ARAP enables international PhD students to spend 1 to 2 years at ASTAR research institutes in Singapore, co-supervised by an ASTAR researcher and their home university supervisor. For those interested in a shorter stay, we have some lab funding to support. 🎯 What we’re looking for International PhD student (or intending to be) with a strong background in AI, or algorithms. A home university supervisor willing to collaborate with A*STAR. ✅ What ARAP covers Monthly stipend of S$3,600. One-time airfare grant. One-time settling-in allowance + housing allowance + medical insurance + conference & IT support. 📅 How to apply If you’re interested, please send me (mile_sikic@a-star.edu.sg) your CV, research statement, and a brief description of your PhD topic. We’ll discuss potential alignment with our lab, and we can nominate you for the ARAP. The deadline is 30th Nov 2025.

We’re opening several postdoctoral and scientist positions focused on RNA structure prediction and mRNA drug optimisation at the @astar_gis. If you’re passionate about AI for therapeutics — especially in areas like generative models, diffusion models, flow matching, or joint embeddings — DM me. 🔗 Open positions: 🔹 AI Scientist, Generative AI Models for the Creation of RNA-Based Vaccines (GIS) - lnkd.in/gRA2mSiA 🔹 Scientist, AI in Drug Discovery (GIS) - lnkd.in/gkASi3Vi #AI #RNA #therapeutics

Do you know ~60% of human SVs fall in ~1% of GRCh38? See our new preprint: arxiv.org/abs/2509.23057 and the companion blog post on how we started this project and longdust: lh3.github.io/2025/09/29/on-…. Work with @QianAlvinQin1

minimap2.com is potentially a phishing site. Please don't use anything from that website. github.com/lh3/minimap2/i…

Published finally! Our study describing the use of metagenomics for AMR/pathogen surveillance in food-centres island-wide in Singapore is out in @npjAMR: nature.com/articles/s4425…

Our heartiest congratulations to our colleagues at A*STAR Genome Institute of Singapore (@astar_gis) who were conferred various awards for their distinguished service and significant contributions to A*STAR and Singapore🇸🇬 🥳 ~ 2025 Fellows of A*STAR: Dr @KokHaoChen & Professor @msikic ~ National Day Awards - The Long Service Medal: Mr Kok Tat Hung & Mr Yeo Chor Soon

Our Systema framework for evaluating genetic perturbation response prediction methods is now out in @NatureBiotech ✨ Predicting cellular responses to genetic perturbations is hard and current metrics often overestimate model performance. Systema helps to evaluate perturbation response prediction methods by focusing on perturbation-specific effects rather than systematic variation 🎯 Kudos to my postdoc Ramon Viñas Torné and everyone who made this possible @wiatrak_maciej @mor_nitzan @teichlab @zoe_piran @LiangzeJ @shawnclarkefan

Our work on direct @nanopore sequencing of non-canonical bases in now out in @NatureComms! Read all about it here: nature.com/articles/s4146… Great collab with @ChewWeiLeong and Hirao labs

I have successfully defended my PhD thesis! Huge thanks to my supervisors @msikic and @xbresson . (It actually happened several weeks ago, but I took some time off after the defence and was avoiding social media - highly recommended!)

Uh oh: deep learning models in genomics lose badly to very simple linear models. Could they have been over-hyped? (new paper by @s_anders_m @const_ae and Wolfgang Huber) nature.com/articles/s4159…