A fast adaptive algorithm for computing whole-genome homology maps

Abstract

MotivationWhole-genome alignment is an important problem in genomics for comparing different species, mapping draft assemblies to reference genomes and identifying repeats. However, for large plant and animal genomes, this task remains compute and memory intensive. In addition, current practical methods lack any guarantee on the characteristics of output alignments, thus making them hard to tune for different application requirements.ResultsWe introduce an approximate algorithm for computing local alignment boundaries between long DNA sequences. Given a minimum alignment length and an identity threshold, our algorithm computes the desired alignment boundaries and identity estimates using kmer-based statistics, and maintains sufficient probabilistic guarantees on the output sensitivity. Further, to prioritize higher scoring alignment intervals, we develop a plane-sweep based filtering technique which is theoretically optimal and practically efficient. Implementation of these ideas resulted in a fast and accurate assembly-to-genome and genome-to-genome mapper. As a result, we were able to map an error-corrected whole-genome NA12878 human assembly to the hg38 human reference genome in about 1 min total execution time and <4 GB memory using eight CPU threads, achieving significant improvement in memory-usage over competing methods. Recall accuracy of computed alignment boundaries was consistently found to be >97% on multiple datasets. Finally, we performed a sensitive self-alignment of the human genome to compute all duplications of length ≥1 Kbp and ≥90% identity. The reported output achieves good recall and covers twice the number of bases than the current UCSC browser’s segmental duplication annotation.Availability and implementation https://github.com/marbl/MashMap

Dashing: fast and accurate genomic distances with HyperLogLog

“…HLL lacks another advantage of MinHash; when Min-Hash is used in conjunction with a reversible hash function, it can be used not only to calculate the relevant set cardinalities but also to report the k-mers common between the sets. This can provide crucial hints when the eventual goal is to map a read to (or near) its point of origin with respect to the reference, as is the goal for tools like MashMap [5].…”

“…Since the release of the seminal Mash tool [1], data sketches such as MinHash have become instrumental in comparative genomics. They are used to cluster genomes from large databases [1], search for datasets with certain sequence content [2], accelerate the overlapping step in genome assemblers [3,4], map sequencing reads [5], and find similarity thresholds characterizing species-level distinctions [6]. Whereas MinHash was originally developed to find similar web pages [7], here it is being used to summarize large genomic sequence collections such as reference genomes or sequencing datasets.…”

Alignment and mapping methodology influence transcript abundance estimation

“…To obtain similar sequences within a reference, we mapped the spliced transcript sequences against a version of the genome where all exon segments were hard-masked (i.e., replaced with N). We performed this mapping using MashMap [20], with a segment size 500 and minimum percent identity of 80%. The sequence similar regions were merged (per-chromosome) using BedTools [45] and concatenated, giving a decoy sequence for each chromosome.…”

“…We also attempt to address one of the failure modes of direct alignment against the transcriptome, compared to spliced alignment to the genome: when a sequenced fragment originates from an unannotated genomic locus bearing sequence similarity to an annotated transcript, it can be falsely mapped to the annotated transcript since the relevant genomic sequence is not available to the method. We describe a procedure that makes use of MashMap [20] to identify and extract such sequence similar decoy regions from the genome. The normal Salmon index is then augmented with these decoy sequences, which are handled in a special manner during mapping and alignment scoring, leading to a reduction in such cases of false mappings.…”

Long transposon-rich centromeres in an oomycete reveal divergence of centromere features in Stramenopila-Alveolata-Rhizaria lineages

“…[77]. Wholegenome alignment was computed by MashMap (https://github.com/marbl/MashMap) employing default settings, and was visualized as a dot plot [78].…”