Using the API

The “AI Folding” api The AI Folding API provides a common interface to AI based protein structure prediction tools. The API currently supports openfold (https://github.com/CyrusBiotechnology/openfold) and alphafold ( https://github.com/deepmind/alphafold#alphafold-output ).

To model a single chain protein with alphafold or openfold

cyrus submit ai-folding input.fasta OpenFold, AlphaFold, and ESMFold. The API runs a post-processing protocol on the results to minimize the output structure into the Rosetta energy function and correct any atomic level errors in sidechain positioning (See Notes for more detail).

Contents

Quickstart

Inputs

Options

Outputs

Notes

References

Release Notes

Quickstart

Model a monomer using AlphaFold

cyrus engine submit ai-folding input.fasta --mode=monomer --ai-tool=

...

alphafold```

Model multiple monomers in parallel using OpenFold

cyrus engine submit ai-folding input.fasta input2.fasta --mode=monomer --ai-tool=openfold

To submit an alphafold SingleSeq job you can set the mode to “singleseq” and specify the number of recycles with the “af-n-recycles” flag:

...

Model a monomer using OpenFold with weights trained by DeepMind (AlphaFold)

Default = OpenFold weights

cyrus engine submit ai-folding input.fasta --mode=monomer --ai-tool=openfold --model-sets=alphafold

Create a model using AlphaFold's SingleSeq mode with 2 recycles

cyrus engine submit ai-folding input.fasta --mode=singleseq --ai-tool=alphafold --af-n-recycles=2

Multiple fasta files can be evaluated in a single job like this:

...

Model a multimer (AlphaFold only)

cyrus engine submit ai-folding input.fasta

...

 --mode=

...

multimer --ai-tool=

...

The Openfold tool can run with either the weights trained by deepmind for alphafold, or weights trained by the AlQuarashi lab for openfold. The openfold trained weights are the default, to use the alphafold weights with the openfold tool, run

cyrus submit ai-folding input.fasta --mode=monomer --ai-tool=openfold --model-sets=alphafold

When modeling a large protein, you can use a larger a100 GPU with --gpu-type=a100 and disable the rosetta relax post-processing step with --run-relax=false

To model a multi-chain protein

cyrus submit input.fasta --mode=multimer --ai-tool=alphafold

API Outputs

...

alphafold

Inputs

FASTA file containing sequence(s) of interest to model.

Options

• --ai-toolAI folding tool to run (openfold, alphafold, esmfold), default = alphafold

• --mode Mode to run with AI tool (monomer, multimer, singleseq), default = monomer

• --model-sets The set of model weights to use with OpenFold (alphafold, openfold), default = openfold (See Notes for more detail)

• --existing-model-data Location of existing model data in GCS, default = null

• --precomputed-alignments Directory path to precomputed alignments that will be upload and used for AlphaFold jobs, default = null

• --run-relax Enable or disable the Rosetta relax phase of post-processing, default = true

• --gpu-type Select the GPU type to use (t4, a100), default = t4 (See Notes for more detail)

Outputs

• predictions (directory) AI tool model predictions

• initial_molprobity_reports (directory) Molprobity report for models output by AI tool

• rosetta_relaxed_models (directory) Rosetta relaxed AI tool models

• final_molprobity_reports (directory) Molprobity report for relaxed models

Notes

Model weight sets

• alphafold - weights trained by DeepMind

• openfold - weights trained by the AlQuarashi Lab for OpenFold

API Post-Processing

The API post-processing protocol consists of the following three steps:

1. Generate a molprobity report for the models output from the AI tool

2. Idealize and relax the models output from the AI tool with rosettaRosetta

3. Generate a molprobity report for the rosetta relaxed models.

Alongside the raw output from the AI tool, the API will produce the following directories:

• initial_molprobity_reports – Molprobity reports from step 1 of the post-processing

• rosetta_relaxed_models – relaxed models from step 2

• final_molprobity_reports -- relaxed models from step 3

...

Modeling large proteins

References

AlphaFold Github

OpenFold Github

ESMFold Github