iiinova — where your digital self keeps evolving

Žiga Avsec @avsecz.bsky.social · 30/01/2026

Further references are also listed in Supplementary information from page 73 onwards. Method's section had to be moved to SI along with the associated references.

000

Reposted by Žiga Avsec

tomwardio.bsky.social @tomwardio.bsky.social · 28/01/2026

Our AlphaGenome paper and models are out today! 🧬 I'm really excited to see how people use the model and build on top of it in their biology research🔬 Huge congratulations to the entire team!👏 Read the full paper and access the model: 📄 Paper: goo.gle/4bXlV6y 💻 Code: goo.gle/4k1xrzI

0225

Žiga Avsec @avsecz.bsky.social · 29/01/2026

Thanks! I don't know actually about which track was most relevant to learn that. My guess would be that each one helps a bit.

020

Žiga Avsec @avsecz.bsky.social · 28/01/2026

Finally, huge congratulations to the amazing team behind this work! 👏

130

Žiga Avsec @avsecz.bsky.social · 28/01/2026

Highlighting two manuscript additions: 1. Data ablation (Supp. Fig 12): RNA-seq + ATAC/DNase proved essential for key evals. h/t @anshulkundaje.bsky.social for the suggestion! 2. Variant Analysis: We found evidence to upgrade a 3’ UTR VUS from Ellingford et al 2022 to "likely pathogenic."

141

Žiga Avsec @avsecz.bsky.social · 28/01/2026

AlphaGenome is out in @nature.com today along with model weights! 🧬 📄 Paper: www.nature.com/articles/s41... 💻 Weights: github.com/google-deepm... Getting here wasn’t a straight path. We discussed the story behind the model, paper & API in the following roundtable: youtu.be/V8lhUqKqzUc

48740

Žiga Avsec @avsecz.bsky.social · 18/07/2025

They were learned during training along with other parameters. See page 31 of the manuscript detailing where they were incorporated.

010

Reposted by Žiga Avsec

EMBL-EBI @ebi.embl.org · 26/06/2025

The first draft of the human genome turns 25 today! 🎂 On 26 June 2000, the UK & US announced the first full map of the 3 billion DNA base pairs that make us human. 🌍 A milestone for humanity, powered by global collaboration. www.youtube.com/watch?v=pnWA... #HumanGenome #ScienceHistory 🖥️🧬

youtube.com

The Human Genome Project at 20 - interview with Ewan Birney

YouTube video by European Bioinformatics Institute - EMBL-EBI

14715

Reposted by Žiga Avsec

Anshul Kundaje @anshulkundaje.bsky.social · 25/06/2025

This a really exciting leap forward for genomic sequence to activity gene regulation models. It is a genuine improvement over pretty much all SOTA models spanning a wide range of regulatory, transcriptional and post-transcriptional processes. 1/

27320

Žiga Avsec @avsecz.bsky.social · 25/06/2025

Thanks to the large consortia like ENCODE, FANTOM5, GTEx, 4DNucleome for generating key training data. And finally, thanks to numerous researchers developing inspiring DNA sequence-to-function models, informative evaluations, and interpretation tools.

1103

Žiga Avsec @avsecz.bsky.social · 25/06/2025

Big thanks to the amazing team! They went above and beyond to get us here. It's a big privilege to work with them. Also thanks to our early testers for their feedback.

150

Žiga Avsec @avsecz.bsky.social · 25/06/2025

This model represents a start, and not the end, of our AlphaGenome named research and release journey aiming to help researchers better understand the genome using AI. We’d love to learn about how you’re using it. You can share projects & feedback in alphagenomecommunity.com forum.

140

Žiga Avsec @avsecz.bsky.social · 25/06/2025

A single model can be more convenient to use and deploy - saying this as someone who worked on a model zoo before. SOTA performance across diverse tasks suggests that fine-tuning it on new data should work quite well. One could also extend it with new assays in the future.

110

Žiga Avsec @avsecz.bsky.social · 25/06/2025

This all-in-one model was achieved by lots of amazing engineering that unlocked training on base-resolution data with long sequence context without increasing the total training compute - training a single model actually used half of what we used for Enformer on the same (older generation) hardware.

140

Žiga Avsec @avsecz.bsky.social · 25/06/2025

I see this model as a fusion of many different DNA sequence-to-function models tackling different steps of gene expression. It also has important new features like splice junction prediction to better capture the complexity of different splicing events.

140

Žiga Avsec @avsecz.bsky.social · 25/06/2025

Excited to launch our AlphaGenome API goo.gle/3ZPUeFX along with the preprint goo.gle/45AkUyc describing and evaluating our latest DNA sequence model powering the API. Looking forward to seeing how scientists use it! @googledeepmind

521982

Reposted by Žiga Avsec

VIB.AI @vibai.bsky.social · 20/11/2024

Fantastic start of the #VIBAIsymposium this morning with a welcome by our scientific director @steinaerts.bsky.social and talks by Julien Gagneur and @avsecz.bsky.social

0387

Žiga Avsec @avsecz.bsky.social · 15/11/2024

Hello world! Refreshing to see my feed full of science again:)

2240