Automatic classification of protein structures using low-dimensional structure space mappings

Abstract

BackgroundProtein function is closely intertwined with protein structure. Discovery of meaningful structure-function relationships is of utmost importance in protein biochemistry and has led to creation of high-quality, manually curated classification databases, such as the gold-standard SCOP (Structural Classification of Proteins) database. The SCOP database and its counterparts such as CATH provide a detailed and comprehensive description of the structural and evolutionary relationships of the proteins of known structure and are widely employed in structural and computational biology. Since manual classification is both subjective and highly laborious, automated classification of novel structures is increasingly an active area of research. The design of methods for automated structure classification has been rendered even more important since the recent past, due to the explosion in number of solved structures arising out of various structural biology initiatives.In this paper we propose an approach to the problem of structure classification based on creating and tessellating low dimensional maps of the protein structure space (MPSS). Given a set of protein structures, an MPSS is a low dimensional embedding of structural similarity-based distances between the molecules. In an MPSS, a group of proteins (such as all the proteins in the PDB or sub-samplings thereof) under consideration are represented as point clouds and structural relatedness maps to spatial adjacency of the points. In this paper we present methods and results that show that MPSS can be used to create tessellations of the protein space comparable to the clade systems within SCOP. Though we have used SCOP as the gold standard, the proposed approach is equally applicable for other structural classifications.MethodsIn the proposed approach, we first construct MPSS using pairwise alignment distances obtained from four established structure alignment algorithms (CE, Dali, FATCAT and MATT). The low dimensional embeddings are next computed using an embedding technique called multidimensional scaling (MDS). Next, by using the remotely homologous Superfamily and Fold levels of the hierarchical SCOP database, a distance threshold is determined to relate adjacency in the low dimensional map to functional relationships. In our approach, the optimal threshold is determined as the value that maximizes the total true classification rate vis-a-vis the SCOP classification. We also show that determining such a threshold is often straightforward, once the structural relationships are represented using MPSS.Results and conclusionWe demonstrate that MPSS constitute highly accurate representations of protein fold space and enable automatic classification of SCOP Superfamily and Fold-level relationships. The results from our automatic classification approach are remarkably similar to those found in the distantly homologous Superfamily level and the quite remotely homologous Fold levels of SCOP. The significance of our results are underlined by the fact that most automa...

Structural Bridges through Fold Space

“…The all- β sandwiches and barrels tend to remain partitioned from each other even at the least stringent probability threshold of 0.5 and are often closer in the landscapes to the α + β community than they are to each other. While the majority of previous visualisations of fold space have noted a four class clustering into SCOP classes [ 17 , 21 , 26 ], one study also saw a division between all- β structures [ 13 ]. However, in this case β -meanders and β -zigzags were found to form the basis for this distinction.…”

“…The process of deriving a global landscape from these data can vary and will inevitably involve assumptions about the nature of the underlying relationships and the extent to which structural alignments can reproduce them. For example, using multi-dimensional scaling or principal component analysis can produce lower dimensional embeddings of an array of similarity scores [ 12 – 21 ]. In these spaces, two or three dimensional maps can be visualised which approximate the similarity between any two structures as closely as possible by their distance on the reduced axes.…”

“…Despite the fact that the above studies construct visualisations of the protein universe using a wide range of different methods, they present a generally consistent picture of this space. In particular, a striking partition between structures based on their secondary structure content is immediately evident [ 12 – 14 , 17 , 21 , 26 , 29 , 32 ]. Broadly speaking, this partition agrees with the class level classifications of SCOP and CATH, which consist of all- α domains, all- β domains and mixed αβ domains.…”