Addressing inaccuracies in BLOSUM computation improves homology search performance

Abstract

BackgroundBLOSUM matrices belong to the most commonly used substitution matrix series for protein homology search and sequence alignments since their publication in 1992. In 2008, Styczynski et al. discovered miscalculations in the clustering step of the matrix computation. Still, the RBLOSUM64 matrix based on the corrected BLOSUM code was reported to perform worse at a statistically significant level than the BLOSUM62.Here, we present a further correction of the (R)BLOSUM code and provide a thorough performance analysis of BLOSUM-, RBLOSUM- and the newly derived CorBLOSUM-type matrices. Thereby, we assess homology search performance of these matrix-types derived from three different BLOCKS databases on all versions of the ASTRAL20, ASTRAL40 and ASTRAL70 subsets resulting in 51 different benchmarks in total. Our analysis is focused on two of the most popular BLOSUM matrices — BLOSUM50 and BLOSUM62.ResultsOur study shows that fixing small errors in the BLOSUM code results in substantially different substitution matrices with a beneficial influence on homology search performance when compared to the original matrices. The CorBLOSUM matrices introduced here performed at least as good as their BLOSUM counterparts in ∼75 % of all test cases. On up-to-date ASTRAL databases BLOSUM matrices were even outperformed by CorBLOSUM matrices in more than 86 % of the times. In contrast to the study by Styczynski et al., the tested RBLOSUM matrices also outperformed the corresponding BLOSUM matrices in most of the cases. Comparing the CorBLOSUM with the RBLOSUM matrices revealed no general performance advantages for either on older ASTRAL releases. On up-to-date ASTRAL databases however CorBLOSUM matrices performed better than their RBLOSUM counterparts in ∼74 % of the test cases.ConclusionsOur results imply that CorBLOSUM type matrices outperform the BLOSUM matrices on a statistically significant level in most of the cases, especially on up-to-date databases such as ASTRAL ≥2.01. Additionally, CorBLOSUM matrices are closer to those originally intended by Henikoff and Henikoff on a conceptual level. Hence, we encourage the usage of CorBLOSUM over (R)BLOSUM matrices for the task of homology search.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-016-1060-3) contains supplementary material, which is available to authorized users.

PFASUM: a substitution matrix from Pfam structural alignments

“…For a list of homologous search results ordered by their E -values, this measure represents the fraction of the correctly found, true positive superfamily relations which remain after cutting the list in order to restrict the number of false positives to a certain amount. We set this threshold to 0.01 errors per query (epq) in concordance to other studies [25, 35]. This effectively restricts the number of false positives found within 100 queries to a single false positive.…”

RBLOSUM performs better than CorBLOSUM with lesser error per query

“…Similarly RBLOSUM matrices were recomputed using the algorithm developed by Styczynski et al [ 4 ] and the respective programs were obtained from http://web.mit.edu/bamel/blosum/revised_blosum.c . CorBLOSUM [ 6 ] matrices were directly obtained from http://www.cbs.tu-darmstadt.de/CorBLOSUM/ . BLOSUM50 and BLOSUM62, the most widely used BLOSUM matrices, were considered in the present study.…”

“…These studies are seen carried out using the initial BLOCKS release version5.0. Hess et al have recently proposed CorBLOSUM [ 6 ] by addressing an inaccuracy in the BLOSUM code. They relied on computing the matrices by changing the datatype threshold from integer to float.…”

“…BLOSUM matrices have been constructed using BLOCKS version 5.0 with 27,102 BLOCKS, whereas the latest updated version 14.3 has 6,739,916 BLOCKS. We read with interest the research article by Hess et al (BMC Bioinform 17:189, 2016 ) on CorBLOSUM, wherein it is argued that an inaccuracy in the BLOSUM code affects the cluster memberships of sequences. They show that replacing the integer based clustering threshold to floating point arguably improves the performances of CorBLOSUM over BLOSUM and RBLOSUM matrices.…”

A Generalized Similarity Metric for Predicting Peptide Binding Affinity

“…Matrices such the PMBEC [27], have been generated based on the two models that produce a minor increase in performance but ultimately are vulnerable to the same factors as their predecessors [11]. In 2008, for example, a miscalculation was discovered in the clustering protocol of the BLOSUM matrix [31]. Despite extensive characterization of the mistake, BLOSUM is still the standard for one of the largest alignment-capable databases available to date, BLAST [11].…”