How to run an AlphaFold+ job

What Alphafold+ does

AlphaFold+ runs AlphaFold 2 structure prediction on protein-sequence FASTA files (or CSV files) and generates an interactive report with confidence metrics. Outputs are computational predictions, not experimentally determined structures. Always inspect confidence metrics before interpreting a model.

How to run an AlphaFold+ job

• Step 1. On the LensAI™ Home Page, click "All Analysis"
  

• Step 2. Locate the AlphaFold+ (AlphaFoldPlus - Metaflow) analysis and click on Run Analysis.
  

• Step 3. Inside the run configuration, select an input file in FASTA format from a Dataset. A CSV file can also be used. It is also possible to select a Dataset with several FASTA/CSV files.
  
  The input must be a portal dataset containing one or more protein-sequence FASTA files. Each FASTA file is treated as one prediction unit. If you submit CSV files, it is important that the column protein_name contains the value/name for all rows.
  
  Input FASTA file requirements
  

  • Use standard protein FASTA format.
  • For a single-chain protein, provide one sequence in the FASTA.
  • For multiple chains in one prediction, include multiple FASTA records in the same FASTA file.
  • Avoid problematic characters in sequence headers. Use simple names with letters, numbers, underscores, or hyphens.
  • Only amino-acid protein sequences are expected. Empty files or files without valid FASTA records will fail or should be rejected.

  Example of a valid monomer FASTA file:
  >Example_protein_A MKTAYIAKQRQISFVKSHFSRQDILDLWIYHTQGYFPDWQNYTPGPGIRYPLTFGWCFKLVPVEPEKVEEANKK
  
  Example of a valid multimer FASTA file:
  >Example_complex_chain_A
MKTAYIAKQRQISFVKSHFSRQDILDLWIYHTQGYFPDWQNYTPGPGIRYPLTFGWCFKLVPVEPEKVEEANKK
>Example_complex_chain_B
GSHMASMTGGQQMGRDLYDDDDKDRWGSELEHHHHHH
  
  Input CSV file requirements
  

  • Must contains the following columns: protein_name, chain_name, chain_sequence.
  • protein_name should be unique per CSV file, chain_name can contains arbitrary values, chain_sequence can only contains protein sequences with standard 1-letter amino acids.

  Example of a valid monomer CSV file:
  protein_name,chain_name,chain_sequence
hGLP1,A,HAEGTFTSDVSSYLEGQAAKEFIAWLVKGRG
  
  Example of a valid multimer CSV file:
  protein_name,chain_name,chain_sequence
hiv1_nef,A,MKTAYIAKQRQISFVKSHFSRQDILDLWIYHTQGYFPDWQNYTPGPGIRYPLTFGWCFKLVPVEPEKVEEANKK
hiv1_nef,B,GSHMASMTGGQQMGRDLYDDDDKDRWGSELEHHHHHH

• Step 4. Define the AlphaFold+ prediction options.

  Option	What to select	Guidance
  Expert prediction setting	Which predictions to relax	Relaxation helps removing non-physical artifacts in the structure predictions.

  

• Step 5. Define the output dataset logical name.
  
  The output dataset will receive the ranked predicted structures, AlphaFold artifacts, and the interactive confidence report. Use a descriptive name that identifies the target protein or complex, the preset, and the run context.

• Step 6. Optionally provide experiment details under the General Run Metadata Inputs.
  
  Recommended metadata includes project, target protein or complex, sequence source, preset rationale, database preset, and any run-specific notes that will help interpret the results later.
• Step 7. Click "Run Analysis" to launch the application.
  
  AlphaFold will run on the LensAI backend. It may take some time before results are available.
• Step 8. Monitor the analysis state and results in the "Analyses history" sidebar.