Cargando…

Spectral Jaccard Similarity: A New Approach to Estimating Pairwise Sequence Alignments

Pairwise sequence alignment is often a computational bottleneck in genomic analysis pipelines, particularly in the context of third-generation sequencing technologies. To speed up this process, the pairwise k-mer Jaccard similarity is sometimes used as a proxy for alignment size in order to filter p...

Descripción completa

Detalles Bibliográficos
Autores principales: Baharav, Tavor Z., Kamath, Govinda M., Tse, David N., Shomorony, Ilan
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Elsevier 2020
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7660437/
https://www.ncbi.nlm.nih.gov/pubmed/33205128
http://dx.doi.org/10.1016/j.patter.2020.100081
_version_ 1783609003107942400
author Baharav, Tavor Z.
Kamath, Govinda M.
Tse, David N.
Shomorony, Ilan
author_facet Baharav, Tavor Z.
Kamath, Govinda M.
Tse, David N.
Shomorony, Ilan
author_sort Baharav, Tavor Z.
collection PubMed
description Pairwise sequence alignment is often a computational bottleneck in genomic analysis pipelines, particularly in the context of third-generation sequencing technologies. To speed up this process, the pairwise k-mer Jaccard similarity is sometimes used as a proxy for alignment size in order to filter pairs of reads, and min-hashes are employed to efficiently estimate these similarities. However, when the k-mer distribution of a dataset is significantly non-uniform (e.g., due to GC biases and repeats), Jaccard similarity is no longer a good proxy for alignment size. In this work, we introduce a min-hash-based approach for estimating alignment sizes called Spectral Jaccard Similarity, which naturally accounts for uneven k-mer distributions. The Spectral Jaccard Similarity is computed by performing a singular value decomposition on a min-hash collision matrix. We empirically show that this new metric provides significantly better estimates for alignment sizes, and we provide a computationally efficient estimator for these spectral similarity scores.
format Online
Article
Text
id pubmed-7660437
institution National Center for Biotechnology Information
language English
publishDate 2020
publisher Elsevier
record_format MEDLINE/PubMed
spelling pubmed-76604372020-11-16 Spectral Jaccard Similarity: A New Approach to Estimating Pairwise Sequence Alignments Baharav, Tavor Z. Kamath, Govinda M. Tse, David N. Shomorony, Ilan Patterns (N Y) Article Pairwise sequence alignment is often a computational bottleneck in genomic analysis pipelines, particularly in the context of third-generation sequencing technologies. To speed up this process, the pairwise k-mer Jaccard similarity is sometimes used as a proxy for alignment size in order to filter pairs of reads, and min-hashes are employed to efficiently estimate these similarities. However, when the k-mer distribution of a dataset is significantly non-uniform (e.g., due to GC biases and repeats), Jaccard similarity is no longer a good proxy for alignment size. In this work, we introduce a min-hash-based approach for estimating alignment sizes called Spectral Jaccard Similarity, which naturally accounts for uneven k-mer distributions. The Spectral Jaccard Similarity is computed by performing a singular value decomposition on a min-hash collision matrix. We empirically show that this new metric provides significantly better estimates for alignment sizes, and we provide a computationally efficient estimator for these spectral similarity scores. Elsevier 2020-07-31 /pmc/articles/PMC7660437/ /pubmed/33205128 http://dx.doi.org/10.1016/j.patter.2020.100081 Text en © 2020 The Authors http://creativecommons.org/licenses/by-nc-nd/4.0/ This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
spellingShingle Article
Baharav, Tavor Z.
Kamath, Govinda M.
Tse, David N.
Shomorony, Ilan
Spectral Jaccard Similarity: A New Approach to Estimating Pairwise Sequence Alignments
title Spectral Jaccard Similarity: A New Approach to Estimating Pairwise Sequence Alignments
title_full Spectral Jaccard Similarity: A New Approach to Estimating Pairwise Sequence Alignments
title_fullStr Spectral Jaccard Similarity: A New Approach to Estimating Pairwise Sequence Alignments
title_full_unstemmed Spectral Jaccard Similarity: A New Approach to Estimating Pairwise Sequence Alignments
title_short Spectral Jaccard Similarity: A New Approach to Estimating Pairwise Sequence Alignments
title_sort spectral jaccard similarity: a new approach to estimating pairwise sequence alignments
topic Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7660437/
https://www.ncbi.nlm.nih.gov/pubmed/33205128
http://dx.doi.org/10.1016/j.patter.2020.100081
work_keys_str_mv AT baharavtavorz spectraljaccardsimilarityanewapproachtoestimatingpairwisesequencealignments
AT kamathgovindam spectraljaccardsimilarityanewapproachtoestimatingpairwisesequencealignments
AT tsedavidn spectraljaccardsimilarityanewapproachtoestimatingpairwisesequencealignments
AT shomoronyilan spectraljaccardsimilarityanewapproachtoestimatingpairwisesequencealignments