Author

John Canny

Google (United States)ORCID 0000-0002-7161-7927Unclaimed ORCID profile · follow to get notified of new preprints

Biology Disciplines

Click to filter author papers

Signature Assays, Organisms & Methods

Click to filter · or open tag

Approach:deep learning· 2 Assay:protein sequence· 2 Research Focus:protein structure prediction· 2 Approach:benchmark study· 1 Research Focus:benchmarking· 1 Assay:protein structure· 1 Research Focus:sequence-to-function modeling· 1 Approach:statistical model· 1

Frequent Co-Authors

Pieter AbbeeliD· 2 shared Roshan RaoiD· 2 shared Alexander W. RivesiD· 1 shared Jason J. LiuiD· 1 shared Joshua MeieriD· 1 shared Neil ThomasiD· 1 shared Nicholas BhattacharyaiD· 1 shared Robert VerkuiliD· 1 shared

Activity by Year

2019-2021 · peak 1/yr

Co-Author Collaboration Network

10 top collaborators · 30 mutual co-author links

Click any collaborator node to inspect or filter shared papers

Focal author (center)Selected co-authorORCID co-authorUnlinked name

Curved chords = mutual co-authorship

2 shared papersSynthetic Biology

Pieter AbbeelORCID

Mutual co-authors in cluster

Author profile →

2 papers

Recent·Most cited

MSA Transformer

Roshan Rao, Jason J. Liu, Robert Verkuil, Joshua Meier, John Canny, Pieter Abbeel, Tom Sercu, Alexander W. Rives

One protein language model trained across many families on whole multiple sequence alignments, with interleaved row and column attention, reportedly outperforms unsupervised structure-learning methods by a wide margin.

bioRxiv · Feb 12, 2021 · PreprintSynthetic Biology

Evaluating Protein Transfer Learning with TAPE

Roshan Rao, Nicholas Bhattacharya, Neil Thomas, Yan Duan, Xi Chen, John Canny, Pieter Abbeel, Yun S. Song

Across five protein tasks with evolution-aware splits, self-supervised pretraining helped nearly every model on every task, yet alignment-based features still beat learned embeddings on secondary structure and contact prediction.

bioRxiv · Jun 20, 2019 · PreprintSynthetic Biology✦ In-depth digest