Recently, there is an excellent structure search tool named Foldseek to identify protein structures sharing similar 3D fold with your query. The work has been published in Nature Biotechnology (2023) (DOI:https://doi.org/10.1038/s41587-023-01773-0) and promoted in Nature (2023). The source code can be obtained at GitHub.
Foldseek claims it searches structures in existing databases including experimental structure database PDB (protein databank) and alphafold predicted structures provided by Deepmind. Foldseek also compared the efficiency with other known software/server such as the Dali server and Combinatorial Extension.
I installed Foldseek in our workstations and tested it with a few known structures. The search speed was amazingly rapid and the output format is very user-friendly in both column-separate txt or HTML.
Here are some example commands and outputs:
Run foldseek to search structures which are similar to 1UBQ (ubiquitin), limited to human (taxonomy id 9606), output as a text file and the alignment scoring is TM
foldseek easy-search pdb/1ubq.pdb /opt/foldseek/afdb --taxon-list 9606 result-human-1UBQ.txt --alignment-type 1 tmp
Run foldseek to search structures which are similar to 1UBQ (ubiquitin), against all alphafold structures, output as a HTML file and the alignment scoring is TM
foldseek easy-search pdb/1ubq.pdb /opt/foldseek/afdb result-all-1UBQ.html --alignment-type 1 --format-mode 3 tmp
Run foldseek to search structures which are similar to 1UBQ (ubiquitin), against all alphafold structures, output as a text file with query, target+header and evalue.
foldseek easy-search pdb/1ubq.pdb /opt/foldseek/afdb result-all-1UBQ.txt --alignment-type 1 --format-output "query,theader,evalue" --format-mode 3 tmp
Usage:
foldseek easy-search <reference structure> <database path> <output file and format> --alignment-type 1 (1 is TM scoring function) --format-mode 3 (3 is HTML format) (remove it for text mode) tmp
The HTML format is very convenient. Click the color bar with the AF code, and one can see the aligned structure (red) compared to the query one (grey).
usage: foldseek easy-search <i:PDB|mmCIF[.gz]> ... <i:PDB|mmCIF[.gz]>|<i:stdin> <i:targetFastaFile[.gz]>|<i:targetDB> <o:alignmentFile> <tmpDir> [options]
options:
-s FLOAT Sensitivity: 1.0 faster; 4.0 fast; 7.5 sensitive [9.500]
--max-seqs INT Maximum results per query sequence allowed to pass the prefilter (affects sensitivity) [1000]
--exhaustive-search BOOL Turns on an exhaustive all vs all search by by passing the prefilter step [0]
--min-seq-id FLOAT List matches above this sequence identity (for clustering) (range 0.0-1.0) [0.000]
-c FLOAT List matches above this fraction of aligned (covered) residues (see --cov-mode) [0.000]
--cov-mode INT 0: coverage of query and target
1: coverage of target
2: coverage of query
3: target seq. length has to be at least x% of query length
4: query seq. length has to be at least x% of target length
5: short seq. needs to be at least x% of the other seq. length [0]
--max-accept INT Maximum accepted alignments before alignment calculation for a query is stopped [2147483647]
-e DOUBLE List matches below this E-value (range 0.0-inf) [1.000E+01]
--seq-id-mode INT 0: alignment length 1: shorter, 2: longer sequence [0]
--alt-ali INT Show up to this many alternative alignments [0]
--num-iterations INT Number of iterative profile search iterations [1]
--tmscore-threshold FLOAT accept alignments with a tmsore > thr [0.0,1.0] [0.000]
--tmalign-hit-order INT order hits by 0: (qTM+tTM)/2, 1: qTM, 2: tTM, 3: min(qTM,tTM) 4: max(qTM,tTM) [0]
--tmalign-fast INT turn on fast search in TM-align [1]
--lddt-threshold FLOAT accept alignments with a lddt > thr [0.0,1.0] [0.000]
--prefilter-mode INT prefilter mode: 0: kmer/ungapped 1: ungapped, 2: nofilter [0]
--alignment-type INT How to compute the alignment:
0: 3di alignment
1: TM alignment
2: 3Di+AA [1]
--cluster-search INT first find representative then align all cluster members [0]
--mask-bfactor-threshold FLOAT mask residues for seeding if b-factor < thr [0,100] [0.000]
--format-mode INT Output format:
0: BLAST-TAB
1: SAM
2: BLAST-TAB + query/db length
3: Pretty HTML
4: BLAST-TAB + column headers
5: Calpha only PDB super-posed to query
BLAST-TAB (0) and BLAST-TAB + column headers (4)support custom output formats (--format-output)
(5) Superposed PDB files (Calpha only) [0]
--format-output STR Choose comma separated list of output columns from: query,target,evalue,gapopen,pident,fident,nident,qstart,qend,qlen
tstart,tend,tlen,alnlen,raw,bits,cigar,qseq,tseq,qheader,theader,qaln,taln,mismatch,qcov,tcov
qset,qsetid,tset,tsetid,taxid,taxname,taxlineage,
lddt,lddtfull,qca,tca,t,u,qtmscore,ttmscore,alntmscore,rmsd,prob
[query,target,fident,alnlen,mismatch,gapopen,qstart,qend,tstart,tend,evalue,bits]
--greedy-best-hits BOOL Choose the best hits greedily to cover the query [0]
--threads INT Number of CPU-cores used (all by default) [16]
-v INT Verbosity level: 0: quiet, 1: +errors, 2: +warnings, 3: +info [3]
--compressed INT Write compressed output [0]
examples:
# Search a single/multiple PDB file against a set of PDB files
foldseek easy-search examples/d1asha_ examples/ result.m8 tmp
# Format output differently
foldseek easy-search examples/d1asha_ examples/ result.m8 tmp --format-output query,target,qstart,tstart,cigar
# Align with TMalign (global)
foldseek easy-search examples/d1asha_ examples/ result.m8 tmp --alignment-type 1
# Skip prefilter and perform an exhaustive alignment (slower but more sensitive)
foldseek easy-search examples/d1asha_ examples/ result.m8 tmp --exhaustive-search 1
references:
- van Kempen, M., Kim, S.S., Tumescheit, C., Mirdita, M., Lee, J., Gilchrist, C.L.M., Söding, J., and Steinegger, M. Fast and accurate protein structure search with Foldseek. Nature Biotechnology, doi:10.1038/s41587-023-01773-0 (2023)
Show an extended list of options by calling 'foldseek easy-search -h'.
KPWu's group research site 