Computational Geneticist @instadeepai and adjunct lecturer at @CayetanoHeredia. Previously at @UGI_at_UCL🇬🇧 and @institutpasteur🇫🇷. Views my own.
Paris, France
Joined February 2013
- Tweets572
- Following623
- Followers757
- Likes1.1K
Pinned Tweet
Our study with @g_hellenthal on detecting adaptive signals before and after European contact in #admixed Latin Americans is finally out in @MolBioEvol
academic.oup.com/mbe/advance…
Javier Mendoza-Revilla retweeted
🚀 Introducing Nucleotide Transformer v3 (NTv3)
Today, we are very excited to share our latest foundation model for biology - Nucleotide Transformer v3 (NTv3).
NTv3 is @instadeepai new multi-species genomics foundation model, designed for 1 Mb, single-nucleotide-resolution prediction, and for bridging representation learning, sequence-to-function modeling, and generative regulatory design within a single framework 🧬
This work was developed in close collaboration with @AlexanderStark8 , @volokuleshov , and @pkoo562 , and reflects several years of joint effort at the intersection of machine learning and regulatory genomics.
Javier Mendoza-Revilla retweeted
Nucleotide Transformer is a series of genomics foundation models of different parameter sizes and training datasets which can be applied to various downstream tasks by fine-tuning. @instadeepai
nature.com/articles/s41592-0…
Javier Mendoza-Revilla retweeted
We were invited to write a review article on DNA/genomic language models (gLMs). We took this occasion to gather our thoughts on promising applications, and major considerations for developing and evaluating gLMs. Pls share with your colleagues: Preprint: arxiv.org/abs/2407.11435
Javier Mendoza-Revilla retweeted
🌱 As scientists wrestled with the mysteries of plant genomics, our #AI problem solvers had a daring question, "Can AI make a difference?" @Nature covered how our #research and AI #genomics teams teamed up to search for an answer, with AgroNT – our new #LLM. 📚Read more→ go.nature.com/3xWopR7
🤗 Access our LLM #open-source on : huggingface.co/InstaDeepAI/a…
Javier Mendoza-Revilla retweeted
A DNA-based large language model, AgroNT, trained on multiple plant genomes, can accurately predict various molecular phenotypes within plant species, including important crops.
@instadeepai @JavierMenRev @thomas_pierrot @LopezMarie05
doi.org/10.1038/s42003-024-0…
Our work on the development of a #DNA Large Language Model trained on genomes from various plants, #AgroNT, is now published in @CommsBio
nature.com/articles/s42003-0…
Javier Mendoza-Revilla retweeted
ChatNT: A Multimodal Conversational Agent for DNA, RNA and Protein Tasks instadeep.com/wp-content/upl…
Javier Mendoza-Revilla retweeted
Very proud to announce ChatNT, the first multimodal conversational agent for biological sequences🧬. ChatNT can be simply prompted with a question and a nucleotide sequence to solve DNA, RNA and Protein tasks 🧑🔬!
📚Paper: tinyurl.com/chatNT-pdf
🌐Blog: tinyurl.com/chatNT-blog
The NT Family is growing! 🐣
✨ Introducing ChatNT, a Conversational Agent designed to analyse genomics sequences and address a wide range of key biological questions, assisting scientists in their daily work 👩🔬
📚Paper: tinyurl.com/chatNT-pdf
🌐Blog: tinyurl.com/chatNT-blog
Javier Mendoza-Revilla retweeted
Glad that our sequence model of promoters in human genome is now published in @ScienceMagazine. Check out the paper for a deep dive into the sequence basis of transcription initiation at the basepair level: science.org/doi/10.1126/scie…
Javier Mendoza-Revilla retweeted
Out now! Together with members of our Exec & Science Leadership Team @emblebi @sangerinstitute @GSK @genentech @pfizer @bmsnews @sanofi we share our perspective on how @OpenTargets we use human genetics & genomics for drug target ID & prioritisation: annualreviews.org/content/jo…
Javier Mendoza-Revilla retweeted
✨ Meet SegmentNT: the first-ever LLM capable of annotating DNA sequences at single nucleotide resolution.
Built on top of our Nucleotide Transformer, SegmentNT offers precise genome annotation surpassing traditional methods and can offer deeper insights into our genome 🧬
Javier Mendoza-Revilla retweeted
InstaDeep's @JavierMenRev will share how AI systems - and LLMs specifically - deepen our understanding of the Maize genome tomorrow at 7:20 pm at the Maize Genetics Meeting in Raleigh, NC. #MGM2024
Read the paper: biorxiv.org/content/10.1101/…
Javier Mendoza-Revilla retweeted
Reposting this criticism of All of Us and their use of ancestry labels
xitter.cf/Graham_Coop/status/162…
The analyses highlighted by the authors risk exacerbating the confusion. Genetic ancestry does not map neatly onto, nor is it highly concordant with these geographic and ethnic categories. 5/n
ALT We evaluated the performance of the ancestry predictions against the self-reported ethnicity of the All of Us samples as ground truth. The performance should be worse than the holdout HGDP samples, but this is expected. Self-reported ethnicity does not correspond to the populations listed above and is prone to false reporting. “Correct” labeling between HGDP/1kg populations and All of Us ethnicities: 1. African (AFR) → Black 2. Latino/Ad Mixed American (AMR) → Hispanic 3. East Asian (EAS) → Asian 4. Finnish (FIN) → White 5. Middle Eastern (MID) → MENA 6. Non-Finnish European (NFE) → White 7. Other (OTH) → Other (do not include skipped) 8. South Asian (SAS) → Asian Based on the procedure above, the concordance between self-reported ethnicity and the ancestry predictions: 0.877
Javier Mendoza-Revilla retweeted
We released the weights and downstream task datasets of the agroNT, our 1B parameters DNA foundation model for plant genomics 🧬🪴. Check out our github (github.com/instadeepai/nucle…) for Jax lovers⚡❤️and our @huggingface space (huggingface.co/InstaDeepAI) for the pytorch version🤗.
Javier Mendoza-Revilla retweeted
I am looking for PhD research interns at the
@instadeepai Paris office for the coming summer. Join us if you are interested in DNA LLMs and regulatory genomics!
We have different projects available and are flexible to find one that fits your PhD.
Link: instadeep.com/internship-off…
Javier Mendoza-Revilla retweeted
We are very excited to present a major breakthrough achievement – the de novo design of synthetic enhancers for selected tissues in fruit fly embryos in vivo using deep- and transfer learning, @deAlmeida_BPet al published today in @Nature nature.com/articles/s41586-0…. Thread 👇(1/N)
Javier Mendoza-Revilla retweeted
If you are interested in gene-environment interactions, molecular phenotypes and single-cell genomics, join us at @institutpasteur in the sunny Paris!
Post-doctoral position in single-cell genomics at Institut Pasteur in Paris docs.google.com/document/d/e…
I am very interested to know what other population and statistical #geneticists think about this.
(1) Pleiotropy: Meaning, a variant at a locus can affect multiple traits simultaneously. Also, pleiotropy can be antagonistic, implying that a variant may be simultaneously harmful AND beneficial. What, then, will these types of screenings be selecting for?