Table 3 The 8 different kinds of MSAs, sequence databases, template databases and parameter settings for AlphaFold.

Name of MSA	MSA generation tool	Sequence database	Template database	AlphaFold2 network/ num_recycles/num_ensemble
default	HHblits, JackHMMER	UniRef30 and BFD, MGnify clusters	pdb70	monomer/8/8
default_seq_temp	HHblits, JackHMMER	UniRef30 and BFD, MGnify clusters	PDB_sort90	monomer/8/8
original	HHblits, JackHMMER	UniRef30, BFD, MGnify clusters	pdb70	monomer/8/8
ori_seq_temp	HHblits, JackHMMER	UniRef30, BFD, MGnify clusters	PDB_sort90	monomer/8/8
colabfold	MMseq2	ColabFold DB	pdb70	monomer/8/8
colab_seq_temp	MMseq2	ColabFold DB	PDB_sort90	monomer/8/8
img	DeepMSA	UniRef90, Integrated Microbial Genomes (IMG), metagenome sequence databases	pdb70	monomer/8/8
img_seq_temp	DeepMSA	UniRef90, Integrated Microbial Genomes (IMG), metagenome sequence databases	PDB_sort90	monomer/8/8

Corresponding to each MSA, a set of structural templates is identified for a target from each of the two structure template databases (pdb70 or PDB_sort90) by using HHsearch to search the MSA against it. default and default_seq_temp (original and ori_seq_temp, colabfold and colab_seq_temp, or img and img_seq_temp) are the same MSA, but their corresponding templates are identified from the different template databases, leading to 8 different combinations of MSAs and templates. In all the cases, the “monomer” network of AlphaFold2 is used to generate structural models. The num_recycles and num_ensemble are set to 8.

Quick links

Search