Simpler protein domain identification using spectral clustering

Abstract

AbstractThe decomposition of a biomolecular complex into domains is an important step to investigate biological functions and ease structure determination. A successful approach to do so is theSPECTRUSalgorithm, which provides a segmentation based on spectral clustering applied to a graph coding interatomic fluctuations derived from an elastic network model.We presentSPECTRALDOM, which makes three straightforward and useful additions toSPECTRUS. For single structures, we show that high quality partitionings can be obtained from a graph Laplacian derived from pairwise interactions–without normal modes. For sets of homologous structures, we introduce a Multiple Sequence Alignment mode, exploiting both the sequence based information (MSA) and the geometric information embodied in experimental structures. Finally, we propose to analyse the clusters/- domains delivered using the so-calledD-family-matching algorithm, which establishes a correspondence between domains yielded by two decompositions, and can be used to handle fragmentation issues.Our domains compare favorably to those of the originalSPECTRUS, and those of the deep learning based methodChainsaw. Using two complex cases, we show in particular thatSPECTRALDOMis the only method handling complex conformational changes involving several sub-domains. Finally, a comparison ofSPECTRALDOMandChainsawon the manually curated domain classificationECODas a reference shows that high quality domains are obtained without using any evolutionary related piece of information.SPECTRALDOMis provided in the Structural Bioinformatics Library, seehttp://sbl.inria.frandhttps://sbl.inria.fr/doc/Spectral_domain_explorer-user-manual.html.