Accurate statistics for local sequence alignment with position-dependent scoring by rare-event sampling

Detecting High Scoring Local Alignments in Pangenome Graphs

Motivation

Increasing amounts of individual genomes sequenced per species motivate the usage of pangenomic approaches. Pangenomes may be represented as graphical structures, e.g. compacted colored de Bruijn graphs, which offer a low memory usage and facilitate reference-free sequence comparisons. While sequence-to-graph mapping to graphical pangenomes has been studied for some time, no local alignment search tool in the vein of BLAST has been proposed yet.

Results

We present a new heuristic method to find maximum scoring local alignments of a DNA query sequence to a pangenome represented as a compacted colored de Bruijn graph. Our approach additionally allows a comparison of similarity among sequences within the pangenome. We show that local alignment scores follow an exponential-tail distribution similar to BLAST scores, and we discuss how to estimate its parameters to separate local alignments representing sequence homology from spurious findings. An implementation of our method is presented, and its performance and usability are shown. Our approach scales sublinearly in running time and memory usage with respect to the number of genomes under consideration. This is an advantage over classical methods that do not make use of sequence similarity within the pangenome.

Availability and implementation

Source code and test data are available from https://gitlab.ub.uni-bielefeld.de/gi/plast.

Supplementary information

Supplementary data are available at Bioinformatics online.

Statistical significance based on length and position of the local score in a model of i.i.d. sequences

Motivation

The local score of a biological sequence analysis is a mathematical tool largely used to analyse biological sequences. Consequently, determining an accurate estimation of its distribution is crucial.

Results

First, we study the accuracy of classical results on the local score distribution in independent and identically distributed model using a Kolmogorov-Smirnov goodness of fit test. Second, we highlight how the length of the segment that realizes the local score improves the classical setting based on local score only. Finally, we study which part of the sequence contributes to the local score.

Contact

mercier@univ-tlse2.fr.

Score distributions of gapped multiple sequence alignments down to the low-probability tail

Assessing the significance of alignment scores of optimally aligned DNA or amino acid sequences can be achieved via the knowledge of the score distribution of random sequences. But this requires obtaining the distribution in the biologically relevant high-scoring region, where the probabilities are exponentially small. For gapless local alignments of infinitely long sequences this distribution is known analytically to follow a Gumbel distribution. Distributions for gapped local alignments and global alignments of finite lengths can only be obtained numerically. To obtain result for the small-probability region, specific statistical mechanics-based rare-event algorithms can be applied. In previous studies, this was achieved for pairwise alignments. They showed that, contrary to results from previous simple sampling studies, strong deviations from the Gumbel distribution occur in case of finite sequence lengths. Here we extend the studies to multiple sequence alignments with gaps, which are much more relevant for practical applications in molecular biology. We study the distributions of scores over a large range of the support, reaching probabilities as small as 10^{-160}, for global and local (sum-of-pair scores) multiple alignments. We find that even after suitable rescaling, eliminating the sequence-length dependence, the distributions for multiple alignment differ from the pairwise alignment case. Furthermore, we also show that the previously discussed Gaussian correction to the Gumbel distribution needs to be refined, also for the case of pairwise alignments.

Pharmacophore Alignment Search Tool (PhAST): Significance Assessment of Chemical Similarity

Previously, we proposed a ligand-based virtual screening technique (PhAST) based on global alignment of linearized interaction patterns. Here, we applied techniques developed for similarity assessment in local sequence alignments to our method resulting in p-values for chemical similarity. We compared two sampling strategies, a simple sampling strategy and a Markov Chain Monte Carlo (MCMC) method, and investigated the similarity of sampled distributions to Gaussian, Gumbel, modified Gumbel, and Gamma distributions. The Gumbel distribution with a Gaussian correction term was identified as the most similar to the observed empirical distributions. These techniques were applied in retrospective screenings on a drug-like dataset. Obtained p-values were adjusted to the size of the screening library with four different methods. Evaluation of E-value thresholds corroborated the Bonferroni correction as a preferred means to identify significant chemical similarity with PhAST. An online version of PhAST with significance estimation is available at http://modlab-cadd.ethz.ch/.

Significance estimation for sequence-based chemical similarity searching (PhAST) and application to AuroraA kinase inhibitors

BACKGROUND

Chemical similarity searching allows the retrieval of preferred screening molecules from a compound database. Candidates are ranked according to their similarity to a reference compound (query). Assessing the statistical significance of chemical similarity scores helps prioritizing significant hits, and identifying cases where the database does not contain any promising compounds.

METHOD

Our text-based similarity measure, Pharmacophore Alignment Search Tool (PhAST), employs pair-wise sequence alignment. We adapted the concept of E-values as significance estimates and employed a sampling technique that incorporates the principle of importance sampling in a Markov chain Monte Carlo simulation to generate distributions of random alignment scores. These distributions were used to compute significance estimates for similarity scores in a preliminary prospective virtual screen for inhibitors of Aurora A kinase.

CONCLUSION

Assessing the significance of compound similarity computed with PhAST allows for a statistically motivated identification of candidate screening compounds. Inhibitors of Aurora A kinase were retrieved from a large compound library.