DPCfam: Unsupervised protein family classification by Density Peak Clustering of large sequence datasets

Abstract

Proteins that are known only at a sequence level outnumber those with an experimental characterization by orders of magnitude. Classifying protein regions (domains) into homologous families can generate testable functional hypotheses for yet unannotated sequences. Existing domain family resources typically use at least some degree of manual curation: they grow slowly over time and leave a large fraction of the protein sequence space unclassified. We here describe automatic clustering by Density Peak Clustering of UniRef50 v. 2017_07, a protein sequence database including approximately 23M sequences. We performed a radical re-implementation of a pipeline we previously developed in order to allow handling millions of sequences and data volumes of the order of 3 TeraBytes. The modified pipeline, which we call DPCfam, finds ∼ 45,000 protein clusters in UniRef50. Our automatic classification is in close correspondence to the ones of the Pfam and ECOD resources: in particular, about 81% of medium-large Pfam families and 80% of ECOD families can be mapped to clusters generated by DPCfam. In addition, our protocol finds more than 14,000 clusters constituted of protein regions with no Pfam annotation, which are therefore candidates for representing novel protein families. These results are made available to the scientific community through a dedicated repository.

Protein family annotation for the Unified Human Gastrointestinal Proteome by DPCfam clustering

“…Results were in line with what we had observed when running DPCfam on UniRef50 [27]; indeed, the majority of the MCs (88.96%) did not fall into any of these categories. We noted that the fraction of disordered MCs was rather small (about 1%).…”

“…In this regime of 'the rich get richer', elements larger in size are usually the more interesting ones. Moreover, as noted in our previous work [27], MCs of smaller size are, on average, of lower quality. Because of this, we decide to focus our downstream analysis on the set of MCs containing at least 50 members.…”

“…Pre-and post-processing steps are needed depending on each specific case. We briefly review each stage of the pipeline, and we refer to the work of Russo et al [27] for further details.…”

“…Supported by a high-performance modular implementation that ensures scaling over large databases, DPCfam consists of a fully automated alternative tool that can improve the annotation of protein datasets. In previous work, we successfully used DPCfam to classify the UniRef50 dataset (version 2017_07), containing about 23 million sequences [27]. As part of this work, 45,000 MCs were generated, of which 30% consisted of protein regions that lacked Pfam annotation.…”

“…The metadata resulting from this comparison is described in Table 1. For an in-depth description of the analysis we refer to Technical Validation and Russo et al [27]. Metaclusters are classified into four groups: equivalent, reduced, extended, and shifted.…”