Improving the quality AlphaFold predicted gH/gL/gD complex structure by MSA optimization

IMPORTANT: Other parameters also affect the quality of the resulting structures (TBD):

• Number of recycles
• Random seeds

• The goal of this project is obtain better AlphaFold models for gH/gL/gD complex, w.r.t. the following scores:

  • ipTM score
  • pTM score
  • PAE score
  • pLDDT score

MSA optimization strategies

Strategies implemented thus far:

Default strategy

graph LR;
    gH/gL/gD_multi_sequence_fasta --> id1["colabfold_search
    Default params"] --> colabfold_batch --> Output;

Loading

Tom's Strategy

graph LR;
  id1["`**DB:** Orthoherpesviridae_WGS`"]  --> jackhmmer --> a3m2multi.sh --> colabfold_batch --> Output;
  id2["`**Query:** gH/gL/gD_multi_sequence_fasta`"] --> jackhmmer

Loading

Strategy #2 - Variable MSA depth using colabfold_search

graph LR;
    gH/gL/gD_multi_sequence_fasta --> id1["colabfold_search"] --> id2["colabfold_batch
    max_msa = 16:32, 32:64, 64:128, 256:512, 512:1024"] --> Output;

Loading

Strategy #3 - MSAs using WGS sequences at various levels of the taxonomic hierarchy

graph LR;
    id1["`**DB:**  Heunggongvirae(kingdom)/Herpesvirales(order)`"]  --> jackhmmer --> a3m2multi.sh --> colabfold_batch --> Output;
    id2["`**Query:** gH/gL/gD_multi_sequence_fasta`"] --> jackhmmer

Loading

Strategy #4 - MAFFT + uniref90/WGS + GUIDANCE

graph LR;
  id1["`**DB:** uniref100`"]  --> jackhmmer --> a3m2multi.sh --> colabfold_batch --> Output;
  id2["`**Query:** gH/gL/gD_multi_sequence_fasta`"] --> jackhmmer

Loading

Strategy #5 - MULTICOM3

graph LR;
  id2["`**Query:** gH/gL/gD_multi_sequence_fasta`"]  --> MULTICOM3 --> Output;
  

Loading

INSTRUCTIONS

Clone this repository

git clone https://github.com/ntnn19/Tom_Topf.git

Prerequisits

• Install Mambaforge (includes mamba)
• Install Singularity

Pipeline Setup

Mamba (Manual)

1. This workflow can be easily setup manually with the given environment file. Install Snakemake and dependencies using the command: mamba env create -f environment.yml
2. Then activate the newly created environment with: mamba activate hsv-1
3. Execute the pipeline with: ./run_workflow.sh <COLABFOLD_WEIGHTS_DIR> E.g.: ./run_workflow.sh /path/to/download_dir Note: The download directory <COLABFOLD_WEIGHTS_DIR> will be created automatically if it does not exist. It should not be a subdirectory in this repository directory. Please specify an absolute path

Workflow