Yo Akiyama
@yoakiyama.bsky.social
MIT EECS PhD student in solab.org Building ML methods to understand and engineer biology
Pinned
Yo Akiyama
@yoakiyama.bsky.social
· Aug 5
Scaling down protein language modeling with MSA Pairformer
Recent efforts in protein language modeling have focused on scaling single-sequence models and their training data, requiring vast compute resources that limit accessibility. Although models that use ...
biorxiv.org
Excited to share work with
Zhidian Zhang, @milot.bsky.social, @martinsteinegger.bsky.social, and @sokrypton.org
biorxiv.org/content/10.1...
TLDR: We introduce MSA Pairformer, a 111M parameter protein language model that challenges the scaling paradigm in self-supervised protein language modeling🧵
Zhidian Zhang, @milot.bsky.social, @martinsteinegger.bsky.social, and @sokrypton.org
biorxiv.org/content/10.1...
TLDR: We introduce MSA Pairformer, a 111M parameter protein language model that challenges the scaling paradigm in self-supervised protein language modeling🧵
Reposted by Yo Akiyama
MMseqs2-GPU sets new standards in single query search speed, allows near instant search of big databases, scales to multiple GPUs and is fast beyond VRAM. It enables ColabFold MSA generation in seconds and sub-second Foldseek search against AFDB50. 1/n
📄 www.nature.com/articles/s41...
💿 mmseqs.com
📄 www.nature.com/articles/s41...
💿 mmseqs.com
GPU-accelerated homology search with MMseqs2 - Nature Methods
Graphics processing unit-accelerated MMseqs2 offers tremendous speedups for homology retrieval from metagenomic databases, query-centered multiple sequence alignment generation for structure predictio...
www.nature.com
September 21, 2025 at 8:06 AM
MMseqs2-GPU sets new standards in single query search speed, allows near instant search of big databases, scales to multiple GPUs and is fast beyond VRAM. It enables ColabFold MSA generation in seconds and sub-second Foldseek search against AFDB50. 1/n
📄 www.nature.com/articles/s41...
💿 mmseqs.com
📄 www.nature.com/articles/s41...
💿 mmseqs.com
Reposted by Yo Akiyama
MMseqs2 v18 is out
- SIMD FW/BW alignment (preprint soon!)
- Sub. Mat. λ calculator by Eric Dawson
- Faster ARM SW by Alexander Nesterovskiy
- MSA-Pairformer’s proximity-based pairing for multimer prediction (www.biorxiv.org/content/10.1...; avail. in ColabFold API)
💾 github.com/soedinglab/M... & 🐍
- SIMD FW/BW alignment (preprint soon!)
- Sub. Mat. λ calculator by Eric Dawson
- Faster ARM SW by Alexander Nesterovskiy
- MSA-Pairformer’s proximity-based pairing for multimer prediction (www.biorxiv.org/content/10.1...; avail. in ColabFold API)
💾 github.com/soedinglab/M... & 🐍
August 5, 2025 at 8:25 AM
MMseqs2 v18 is out
- SIMD FW/BW alignment (preprint soon!)
- Sub. Mat. λ calculator by Eric Dawson
- Faster ARM SW by Alexander Nesterovskiy
- MSA-Pairformer’s proximity-based pairing for multimer prediction (www.biorxiv.org/content/10.1...; avail. in ColabFold API)
💾 github.com/soedinglab/M... & 🐍
- SIMD FW/BW alignment (preprint soon!)
- Sub. Mat. λ calculator by Eric Dawson
- Faster ARM SW by Alexander Nesterovskiy
- MSA-Pairformer’s proximity-based pairing for multimer prediction (www.biorxiv.org/content/10.1...; avail. in ColabFold API)
💾 github.com/soedinglab/M... & 🐍
Reposted by Yo Akiyama
Side story: While working on the Google Colab notebook for MSA pairformer. We encountered a problem: The MMseqs2 ColabFold MSA did not show any contacts at protein interfaces, while our old HHblits alignments showed clear contacts 🫥... (2/4)
August 5, 2025 at 7:39 AM
Side story: While working on the Google Colab notebook for MSA pairformer. We encountered a problem: The MMseqs2 ColabFold MSA did not show any contacts at protein interfaces, while our old HHblits alignments showed clear contacts 🫥... (2/4)
Excited to share work with
Zhidian Zhang, @milot.bsky.social, @martinsteinegger.bsky.social, and @sokrypton.org
biorxiv.org/content/10.1...
TLDR: We introduce MSA Pairformer, a 111M parameter protein language model that challenges the scaling paradigm in self-supervised protein language modeling🧵
Zhidian Zhang, @milot.bsky.social, @martinsteinegger.bsky.social, and @sokrypton.org
biorxiv.org/content/10.1...
TLDR: We introduce MSA Pairformer, a 111M parameter protein language model that challenges the scaling paradigm in self-supervised protein language modeling🧵
Scaling down protein language modeling with MSA Pairformer
Recent efforts in protein language modeling have focused on scaling single-sequence models and their training data, requiring vast compute resources that limit accessibility. Although models that use ...
biorxiv.org
August 5, 2025 at 6:31 AM
Excited to share work with
Zhidian Zhang, @milot.bsky.social, @martinsteinegger.bsky.social, and @sokrypton.org
biorxiv.org/content/10.1...
TLDR: We introduce MSA Pairformer, a 111M parameter protein language model that challenges the scaling paradigm in self-supervised protein language modeling🧵
Zhidian Zhang, @milot.bsky.social, @martinsteinegger.bsky.social, and @sokrypton.org
biorxiv.org/content/10.1...
TLDR: We introduce MSA Pairformer, a 111M parameter protein language model that challenges the scaling paradigm in self-supervised protein language modeling🧵