Tutorial 3.5 Prediction parameters

Introduction

LIT-AlphaFold allows to easily modify many of the parameters used for structure prediction during calculation. In this tutorial all the available prediction parameters are described.

General parameters

Most of the parameters used during structure prediction are written in a the config.yaml file. Between them the most important parameter is the run parameters, which indicates which prediction protocol should be used during calculation. Information regarding the specific prediction protocols are saved in multiple configuration files, one for each method.

hydra:
  run:
    dir: hydra_prediction/${now:%Y-%m-%d}/${now:%H-%M-%S}

defaults:
  - run: default
  - _self_

weights:
  data_dir: /alpha/alphafold_data
  multimer_type: v3
  monomer_type: _ptm


mode: 'custom' # 'custom', 'pulldown', 'all_vs_all', 'homo-oligomer'
output_path: Protein_structures
monomer_objects_dir: ???
input_file: ???
save_multimers: False
compress_multimers: False
benchmark: False
logger_file: alphafold_prediction
optimize: False

save_all: False
models_to_relax: null #'best', 'all'
allow_resume: True
remove_unpaired_msa: False
shuffle_templates: False
multi_monomer: False

mutate_msa_file: null
remove_msa_region: null
modify_unpaired_msa: True
modify_paired_msa: False

A description of some of the most important parameters is given below.

Allow resume

Allows AlphaFold to continue the calculations from the latest saved result in case a previous calculation failed.

Save all

AlphaFold and AlphaFold-Multimer generate a large number of outputs. To save space by default certain results are not saved to disk. By enabling save_all additional information are saved: distogram, masked_msa, experimentally resolved weights.

Model to relax

Indicates for which model minimization should be performed.

Multi monomer

Use AlphaFold-Multimer to perform structure predictions for monomeric proteins, in this case AlphaFold2 will not be used to run predictions.

Run Parameters

Certain parameters are modified by various protocols aiming at sampling alternative protein conformations or multimer interfaces. For each protocol a specific configuration file has been prepared with preset parameters, to facilitate their use.

num_cycle_mono: 5
num_cycle_multi: 20
max_seq: null
max_extra_seq: null
num_predictions_per_model: 1
dropout: False
cluster_profile: True
only_template: False
use_templates: True
remove_template_msa: False
multimer_templates: null #replace with 'all' or 'inter'
remove_pair_msa: False

A description of some of the most important parameters is given below.

Number of cycles

The arguments num_cycle_mono and num_cycle_multi describe the number of recycle steps used during calculation. Increasing the number of recycles can lead to better structures.

For multimer predictions there is an early stopping conditions, which interrupts the calculation once the structure has converged to a minima.

Number of sequences

Number of MSA sequences and MSA extra sequences used during calculations, it overrides the default values. Subsampling of the MSA has been used to obtain a more diverse conformational ensembles.

Number of prediction per model

AlphaFold and AlphaFold-Multimer are formed by five models, each one generating a different structure. Currently it is suggested for monomers to perform calculations once, while for multimers it is suggested to repeat them five times.

Multiple articles have shown that repeating calculations a large number of times can be beneficial, both for conformational sampling, and to obtain better multimer structures.

Dropout

Dropout is a technique used for the training of deep learning networks. It conists in randomly dropping some of the nodes to force the model to learn redundant solutions for the objective function. By applying dropout at inference time the network weights are perturbed, leading to more diverse outputs. This approach has been applied to obtain a better conformational sampling for multimer structure prediction.

Cluster profile

See description in ColabFold

Only templates

Limit predictions to models using template information

Multimer templates

Allow AlphaFold multimer to use inter-chain distances from structural templates to perform its predictions.

Remove pair MSA

Removed the paired part of the MSA during AlphaFold-Multimer calculations.

Conclusion

In this tutorial you have learned how to change parameters for structure prediction using LIT-AF.