Yoshitaka Moriwaki

Yoshitaka Moriwaki

@agsmith.bsky.social

とある文京区のJKラボォです。タンパク質の立体構造とバイオインフォマティクスに興味があります。I'm interested in protein bioinformatics. ORCID: 0000-0003-0448-9790

340 followers266 following86 posts

Yoshitaka Moriwaki
@agsmith.bsky.social

Here is a movie of the conformational sampling with RGI-toolkit.

Yoshitaka Moriwaki@agsmith.bsky.social

[6/9] Beyond correcting stereochemistry, RGI-toolkit also enables targeted sampling of multiple protein conformations, a task that remains challenging for AF3. Here, we show an example of a membrane transporter that adopts both outward-facing and inward-facing conformations.

Yoshitaka Moriwaki
@agsmith.bsky.social

[8/9] Simply enable "use_rgi" and specify the target configuration to correct molecular stereochemistry. You can also define custom restraints, including distance restraints, to guide structure prediction.

Yoshitaka Moriwaki
@agsmith.bsky.social

[9/9] We hope RGI-toolkit will help researchers predict structures containing complex natural products and drug candidates with greater stereochemical accuracy, opening up new possibilities for structure-based research!

Yoshitaka Moriwaki
@agsmith.bsky.social

[6/9] Beyond correcting stereochemistry, RGI-toolkit also enables targeted sampling of multiple protein conformations, a task that remains challenging for AF3. Here, we show an example of a membrane transporter that adopts both outward-facing and inward-facing conformations.

Yoshitaka Moriwaki
@agsmith.bsky.social

[4/9] Here is an example: With RGI enabled, the ligand stereochemistry was correctly reproduced as specified by the user. Left: RGI ON / Right: RGI OFF White: Crystal structure (ground truth)

Yoshitaka Moriwaki
@agsmith.bsky.social

[5/9] RGI-toolkit is a unified framework that extends restraint-guided inference (RGI) to six additional structure prediction models beyond Boltz. This enables direct comparisons of predicted structures across different model architectures and trained weights.

Yoshitaka Moriwaki
@agsmith.bsky.social

[3/9] To address this issue, we previously developed a method that continuously applies interatomic restraints during the diffusion process, as reported in our two earlier papers. ACS Omega: doi.org/10.1021/acso... JCTC: doi.org/10.1021/acs....

doi.orgImproving Stereochemical Limitations in Protein–Ligand Complex Structure PredictionAbstract. AlphaFold3 has revolutionized biology by enabling the prediction of protein complexes with various biomolecules, including small molecular ligand

Yoshitaka Moriwaki reposted

Milot Mirdita
@milot.bsky.social

ColabFold 1.6.3 is out! 2.5x faster, pip-installable, ipSAE+pDockQ2 scores. Thanks Choonghwan Lee, Marielle Russo, Gyuri Kim 🐍pip install colabfold[alphafold] CF2 Sneak Peak with AF3/Boltz/Protenix/ESMFold2… 🐍pip install "colabfold[alphafold3]@git+https://github.com/sokrypton/ColabFold@af3-preview"

Yoshitaka Moriwaki
@agsmith.bsky.social

Our work is published. Hori, T., Moriwaki, Y., Ishitani, R.*, Distance-Restraint-Guided Diffusion Models for Sampling Protein Conformational Changes and Ligand Dissociation Pathways, pubs.acs.org/doi/10.1021/...

pubs.acs.orgDistance-Restraint-Guided Diffusion Models for Sampling Protein Conformational Changes and Ligand Dissociation PathwaysProtein conformational dynamics and ligand binding processes are fundamental to biological function, yet their systematic sampling and thermodynamic characterization remain challenging. Here, we present a distance-restraint-guided inference method that extends AlphaFold3-like diffusion model frameworks to predict protein structures at specified conformational states. By restraining intergroup distances─defined as geometric centroid distances between atom groups─during the reverse diffusion process, our method enables systematic sampling along reaction coordinates without model retraining. We implemented this approach in Boltz-2 and demonstrated its effectiveness on three model proteins that undergo open-closed conformational transitions, as well as on a protein-peptide dissociation pathway. Compared with conventional approaches that induce conformational diversity by manipulating the input multiple sequence alignments, our method achieved more uniform coverage of conformational space while maintaining high structural quality as assessed by both learning-based confidence metrics and stereochemistry-based validation. By combining distance-restrained sampling with molecular dynamics simulations, we constructed free energy landscapes and quantitatively estimated binding free energies. Altogether, our approach bridges deep learning-based structure prediction and physics-based simulations, providing an efficient strategy for characterizing the dynamics of biomolecules.

Yoshitaka Moriwaki reposted

日経サイエンス
@nk-science.nikkei.com

【津波情報 2026年4月20日】 津波警報を発表しました。ただちに避難してください。 ［震源に関する情報］ 16時52分頃、三陸沖を震源とする地震。震源の深さは約10km、地震の規模はM7.4と推定。

Yoshitaka Moriwaki reposted

EMBL-EBI
@ebi.embl.org

You asked, we listened. Millions of AI-predicted protein complex structures are now available in the #AlphaFold Database. This spans homodimers from 20 of the most studied species, including humans, as well as the World Health Organization’s priority pathogens list. www.ebi.ac.uk/about/news/t...

Yoshitaka Moriwaki reposted

Milot Mirdita
@milot.bsky.social

My time in @martinsteinegger.bsky.social's group is ending, but I’m staying in Korea to build a lab at Sungkyunkwan University School of Medicine. If you or someone you know is interested in molecular machine learning and open-source bioinformatics, please reach out. I am hiring! mirdita.org

mirdita.orgMirdita Lab - Laboratory for Computational Biology & Molecular Machine LearningMirdita Lab builds scalable bioinformatics methods.

Yoshitaka Moriwaki reposted

Martin Steinegger 🇺🇦
@martinsteinegger.bsky.social

FoldMason is out now in @science.org. It generates accurate multiple structure alignments for thousands of protein structures in seconds. Great work by Cameron L. M. Gilchrist and @milot.bsky.social. 📄 www.science.org/doi/10.1126/... 🌐 search.foldseek.com/foldmason 💾 github.com/steineggerla...

science.orgMultiple protein structure alignment at scale with FoldMasonProtein structure is conserved beyond sequence, making multiple structural alignment (MSTA) essential for analyzing distantly related proteins. Computational prediction methods have vastly extended ou...
Yoshitaka Moriwaki
@agsmith.bsky.social

I submitted this manuscript to a famous journal, but after 2 months, I received an e-mail saying that the editor who originally assigned to it became unavailable and sufficient reviewers have not been assigned yet. I have finally withdrawn this submission and am considering to another journal🤯

Yoshitaka Moriwaki@agsmith.bsky.social

Our new preprint is out. Our group performed a comprehensive protein–protein complex prediction within 2,437 biosynthetic gene clusters. We predicted a total of 487,828 complexes for known BGCs, identifying 15,438 heteromeric interactions with an ipTM ≥ 0.6. (2/3) www.biorxiv.org/content/10.1...

biorxiv.orgPredicting protein complexes in biosynthetic gene clustersBiosynthetic gene clusters (BGCs) are contiguous genomic regions that encode diverse, non-homologous proteins required for the production of specific natural products. Their genetic diversity underlie...
Yoshitaka Moriwaki
@agsmith.bsky.social

The Terada and Leow groups have developed ColabReaction, a cloud-based tool that accelerates transition-state searches using the machine learning potential UMA. This work was featured as a journal cover. available at: colabreaction.net pubs.acs.org/doi/10.1021/...

Yoshitaka Moriwaki@agsmith.bsky.social

The Terada group, in collaboration with Dr. Leow, has developed ColabReaction — a cloud-based tool for accelerated transition state searches using Direct MaxFlux (DMF) and the ML potential (UMA) available at: colabreaction.net Preprint: chemrxiv.org/engage/chemr...

colabreaction.netUnder construction
Yoshitaka Moriwaki
@agsmith.bsky.social

This paper has been published in ACS Omega: pubs.acs.org/doi/10.1021/... Available at Google Colaboratory: github.com/cddlab/colab...

pubs.acs.orgImproving Stereochemical Limitations in Protein–Ligand Complex Structure PredictionAlphaFold3 has revolutionized biology by enabling the prediction of protein complexes with various biomolecules, including small molecular ligands. However, the ligand structures predicted by the mode...
Yoshitaka Moriwaki@agsmith.bsky.social

Our latest work has been published! Our additional implementation for Boltz-1 largely improves the stereochemical accuracy of predicted ligand molecule without retraining the model. We also developed a notebook on Google Colaboratory, colab.research.google.com/github/cddla... .

Yoshitaka Moriwaki reposted

Mohammed AlQuraishi
@moalquraishi.bsky.social

OpenFold3-preview (OF3p) is out: a sneak peek of our AF3-based structure prediction model. Our aim for OF3 is full AF3-parity for every modality. We now believe we have a clear path towards this goal and are releasing OF3p to enable building in the OF3 ecosystem. More👇

Yoshitaka Moriwaki
@agsmith.bsky.social

Our new preprint is out. Our group performed a comprehensive protein–protein complex prediction within 2,437 biosynthetic gene clusters. We predicted a total of 487,828 complexes for known BGCs, identifying 15,438 heteromeric interactions with an ipTM ≥ 0.6. (2/3) www.biorxiv.org/content/10.1...

biorxiv.orgPredicting protein complexes in biosynthetic gene clustersBiosynthetic gene clusters (BGCs) are contiguous genomic regions that encode diverse, non-homologous proteins required for the production of specific natural products. Their genetic diversity underlie...
Yoshitaka Moriwaki
@agsmith.bsky.social

Among them, 1,390 protein pairs exhibited structural homology with an RMSD ≤ 2.0 Å. These predictions highlight intriguing molecular mechanisms involving proteins previously annotated as uncharacterized or potentially dysfunctional. (2/3)

Yoshitaka Moriwaki
@agsmith.bsky.social

We used the ipSAE metric developed by @rolanddunbrack.bsky.social ( github.com/DunbrackLab/... ) alongside ipTM to evaluate the predicted complexes. We found that ipSAE showed greater discriminative power than ipTM in identifying true binding pairs among highly similar proteins. (3/3)

github.comGitHub - DunbrackLab/IPSAE: Scoring function for interprotein interactions in AlphaFold2 and AlphaFold3Scoring function for interprotein interactions in AlphaFold2 and AlphaFold3 - DunbrackLab/IPSAE
Yoshitaka Moriwaki
@agsmith.bsky.social

Now DSSP 4.5.6 and its python module are available on Homebrew. `brew install brewsci/bio/dssp`.

Yoshitaka Moriwaki@agsmith.bsky.social

> Since version 4.5.2 it is possible to build a Python module to directly work with DSSP info inside your Python scripts. github.com/PDB-REDO/dssp It can be installed using `brew install brewsci/bio/dssp`. I hope it will also be available on PyPI.

github.comGitHub - PDB-REDO/dssp: Application to assign secondary structure to proteinsApplication to assign secondary structure to proteins - PDB-REDO/dssp
Yoshitaka Moriwaki
@agsmith.bsky.social

> Participants who consumed alcohol had significantly better observer-ratings for their Dutch language, specifically better pronunciation, compared with those who did not consume alcohol. However, alcohol had no effect on self-ratings of Dutch language skills.

Yoshitaka Moriwaki reposted

ぬまがさワタリ@『いきものニュース図解』ほか3/19同時発売
@numagasa.bsky.social

「科学者はもはやTwitterが仕事の役に立たないと感じ、Blueskyに切り替えた」という新しい研究結果。 過去10年、Twitterは科学者の拠点だったが、マスクの買収によって疑似科学、陰謀論、嫌がらせが蔓延し、多くの科学者がTwitter(X)を見限って、Blueskyに移動。 この現象について改めて研究が行われ、実際「科学者はかつてTwitterから得ていた専門的な利益を、今はBlueskyでより効果的に得ている」というデータが示されている。 科学が弾圧される今、科学者フレンドリーなSNSとしてブルスカの成長に期待 academic.oup.com/icb/advance-...

academic.oup.comScientists no Longer Find Twitter Professionally Useful, and have Switched to BlueskySynopsis. Social media has become widely used by the scientific community for a variety of professional uses, including networking and public outreach. For
Yoshitaka Moriwaki
@agsmith.bsky.social

Boltz-1の開発グループはDiffdockと同じグループなんだけど、なんでかDiffdockと性能比較していないね　あとよく言われるAutodockとも。 するまでもなく性能良くなっているんだろうけれど、論文の査読者的には「なぜ比較していない？」って言うのかな

Yoshitaka Moriwaki reposted

Pedro Beltrao
@pedrobeltrao.bsky.social

Unexpectedly, @jurgjn.bsky.social found that running Alphafold3 predictions for protein interactions can yield ipTM scores that are more predictive of true interactions when run in pools of proteins instead of pairwise predictions. Presumably, this reflects some sort of "competition effect".

Yoshitaka Moriwaki
@agsmith.bsky.social

うーん、足の指を骨折してしまったかもしれん。ただの捻挫かと思ってたけど、すごく紫色に腫れてて痛い 外回りが増えてるから安静にできないのは困るな

Yoshitaka Moriwaki reposted

Spyros Lytras
@spyroslytras.bsky.social

The CoVFit paper is now published @natcomms.nature.com !! Exciting work led by Jumpei Ito really showing the potential of pLMs for genotype-2-phenotype predictions!! www.nature.com/articles/s41...

nature.comA protein language model for exploring viral fitness landscapes - Nature CommunicationsIto et al. present CoVFit, an AI model that predicts variant fitness (transmissibility) from spike protein sequences alone. They further demonstrate its utility in forecasting viral evolution via sing...
Yoshitaka Moriwaki
@agsmith.bsky.social

大学に長年いると、数は少ないながらも実家暮らしの学生の中には、学部の選択、研究室、就職先の企業・業種に至るまで、親の意向に抗えず言いなりの人生を歩んでる子の話をそこそこ聞くことがある 大学入試かつ医学部ならまだ聞く話だが、それ以外でもあるのか、みたいな

Yoshitaka Moriwaki
@agsmith.bsky.social

Our latest work has been published! Our additional implementation for Boltz-1 largely improves the stereochemical accuracy of predicted ligand molecule without retraining the model. We also developed a notebook on Google Colaboratory, colab.research.google.com/github/cddla... .

bioRxiv Bioinfo@biorxiv-bioinfo.bsky.social

Improving Stereochemical Limitations in Protein-Ligand Complex Structure Prediction https://www.biorxiv.org/content/10.1101/2025.03.25.645362v1

Yoshitaka Moriwaki
@agsmith.bsky.social

なんか、引き受けるかどうか熟考を要するような無報酬の仕事依頼メールが1日に何個かくるようになって、本業が進みにくくなっているのを感じる

Yoshitaka Moriwaki
@agsmith.bsky.social

相手が営利企業からだったら一蹴しているけど、学術機関とか学生とか教員だと「うーん……暇だったら引き受けてもいいんだけどな……依頼が多いんよな……うーん」って感じ 最近はSADOKUがやけに回ってくる

Yoshitaka Moriwaki
@agsmith.bsky.social

最近の選挙で、投票所の記名台近くで一人で投票券3枚持った写真が存在するのおかしくね？みたいな騒動あったけど、投票券を偽造して複数枚持ち込むって、そんなに難しいことではないように思えるから（本人確認もないし）、すでに組織的にやられてるんでは？と思った

Yoshitaka Moriwaki
@agsmith.bsky.social

・常に2秒で判断 ・ホワイトボードの前に立つ ・難しい問題をシンプルに説明 ができる学生は確かに前職に何人かいました。皆パワポ資料が徹底的に作り込まれていて舌を巻くレベルだった。今はそういうのが持て囃されている時代なので、それはそれでいいんですけれど、肝心の話の中身は全くもって薄っぺらいのも特徴でした。ちょっと疑問に思ったことを質問すると、その学生は詰まらずさらっと答えているように見せかけて、全然質問の答えになっていない回答になっていることに気付いていないとか。

Yoshitaka Moriwaki
@agsmith.bsky.social

修士の2年間で1ヶ月くらいしか真面目に取り組んでいなかったそんな学生の例が極まっていて、「俺何でもすぐ理解できてプログラミングもバッチリこなせますよ！」「ところでエディタとかfor文ってなんですっけ？」なところから始まった学生がやった研究の内容はあまりにもレベルが低かったけれど、そのスライドはびっくりするくらいわかりやすく整理されてて、ただ既存のデータを整理しただけで凄そうに図表を並べて自信満々に示していた（※新規性ほぼなし）が、なんと修論審査した先生たちからは全員満点ついていたとか（普通そんなことあまりない）。