SIFTS: updated Structure Integration with Function, Taxonomy and Sequences resource allows 40-fold increase in coverage of structure-based annotations for proteins

Abstract

The Structure Integration with Function, Taxonomy and Sequences resource (SIFTS; http://pdbe.org/sifts/) was established in 2002 and continues to operate as a collaboration between the Protein Data Bank in Europe (PDBe; http://pdbe.org) and the UniProt Knowledgebase (UniProtKB; http://uniprot.org). The resource is instrumental in the transfer of annotations between protein structure and protein sequence resources through provision of up-to-date residue-level mappings between entries from the PDB and from UniProtKB. SIFTS also incorporates residue-level annotations from other biological resources, currently comprising the NCBI taxonomy database, IntEnz, GO, Pfam, InterPro, SCOP, CATH, PubMed, Ensembl, Homologene and automatic Pfam domain assignments based on HMM profiles. The recently released implementation of SIFTS includes support for multiple cross-references for proteins in the PDB, allowing mappings to UniProtKB isoforms and UniRef90 cluster members. This development makes structure data in the PDB readily available to over 1.8 million UniProtKB accessions.

Integrated evolutionary and structural analysis reveals xenobiotics and pathogens as the major drivers of mammalian adaptation

“…In order to identify residues that were under positive selection in mammalian evolution, we obtained coding sequences for Eutherian mammals from Ensembl and phylogenetic trees from the Ensembl Compara database 24 . 3D structures corresponding to human proteins in our dataset were then downloaded from PDB 25 and mapped against protein sequences using the SIFTS resource 26 . We aligned coding sequences corresponding to each tree using the PRANK aligner 27 and used the Slr software 28 to detect positively selected sites.…”

“…In rare cases where more than one human protein with a 514 structure was present for an alignment, one was retained at random. Individual residues were 515 then mapped using the SIFTS database 26 . SIFTS provides a mapping between PDB 25 and 516 UniProt 94 sequences and, as the UniProt protein sequences can vary from those in Ensembl,…”

“…In rare cases where more than one human protein with a structure was present for an alignment, one was retained at random. Individual residues were then mapped using the SIFTS database 26 . SIFTS provides a mapping between PDB 25 and UniProt 92 sequences and, as the UniProt protein sequences can vary from those in Ensembl, we performed an additional mapping step by constructing pairwise alignments between UniProt and Ensembl sequences, resulting in a sitewise mapping between Ensembl and PDB residues.…”

The functional landscape of the human phosphoproteome

“…Structural data were collected from the best human experimentally resolved structures contained in the Protein Data Bank (wwPDB consortium 2019) or, alternatively, from the best homology model derived from ModPipe version 2.2.0 (Pieper et al 2014) . Structural coordinates were remapped to UniProt positions using the SIFTS database (Dana et al 2019) . For the 8,870 proteins with structural data, the impact of all possible variants was computed using FoldX v 4.0 (Guerois, Nielsen, and Serrano 2002) .…”

Uncovering of cytochrome P450 anatomy by SecStrAnnotator

“…Set-NR: A list of protein domains annotated as Cytochrome P450 was obtained from SIFTS resource (Dana et al, 2019) via PDBe REST API (available at http://www.ebi.ac.uk/pdbe/api/mappings/:accession) on 23 March 2020. More specifically, annotations originating from databases CATH and Pfam (accessions 1.10.630.10 and PF00067) were merged to obtain 1807 protein domains located in 973 PDB entries.…”