AlphaFold2 using MMseqs2 and WebGPU

Prediction input

59 residues

Job name Alignment

Protein sequence Use : to specify inter-protein chainbreaks for modeling complexes (supports homo- and hetero-oligomers). Template structure (optional) Choose a PDB or mmCIF file Folds the query against a structure you provide.

Advanced settings

Recycles Random seed Clustered MSA rows Extra MSA rows Monomer model manifest URL Used automatically for single-chain and monomer-A3M inputs. Multimer model manifest URL Used automatically for colon-separated chains and ColabFold complex A3Ms.

Removes persistent model shards and the in-memory model. They will be verified and downloaded again on the next run.

MMseqs2 API URL Used only for MMseqs2 MSA mode. The default is ColabFold's public API.

Prediction results

Mean pLDDT—Confidence

pTM—Predicted TM-score

Sequence—Residues

MSA depth—Input rows

3D structure

Colour

Structure viewer loads after prediction.

Predicted lDDT

MSA sequence coverage

Predicted aligned error

Recycle summary

Pass	pLDDT	pTM	Time

Reference

AlphaFold2: Jumper J, Evans R, Pritzel A, et al. Highly accurate protein structure prediction with AlphaFold. Nature, 2021.

MMseqs2 MSA generation is provided by the public ColabFold service. Mirdita M, Schütze K, Moriwaki Y, et al. ColabFold: making protein folding accessible to all. Nature Methods, 2022. Marv, the mascot on this page, is ColabFold’s, under its MIT licence.