ParaDISM: precise mapping of short reads to genes with highly homologous regions

Abstract Background Genes with highly similar genomic copies (paralogs, tandem duplications and pseudogenes) pose a major challenge for Short-Read High Throughput Sequencing (srHTS). High sequence similarity makes it difficult to unambiguously identify the sequences of origin of short reads. This results in misalignment artifacts which can propagate through bioinformatic pipelines and increase error rates in subsequent steps such as single nucleotide variant calling. Results We present ParaDISM , a pipeline that refines standard alignments to improve read placement and reduce misalignment-driven false variant calls in highly homologous sequences. The pipeline uses a multiple sequence alignment of reference sequences to identify disambiguating positions and assign a read/read pair to a sequence only when supported by unambiguous sequence-specific evidence. ParaDISM focuses on specificity and precision (positive predictive value) rather than sensitivity by adopting a conservative strategy of read placement and reporting reads with insufficient evidence for the gene of origin as unassigned. An optional iterative refinement procedure calls variants from confidently assigned reads, updates the reference sequences to gain sample-specific evidence for read placement, and re-processes remaining non-assigned reads to mitigate the loss of sensitivity caused by the stringent criteria for read placement. We evaluated the performance of ParaDISM both in terms of read alignment and the resulting short variant calls using extensive computational simulation experiments and the Genome in a Bottle HG002 benchmark. We applied ParaDISM to reanalyze two case studies: five public tumour exomes at the GNAQ/GNAQP1 locus, and 18 short-read sequencing datasets of patients diagnosed with Autosomal Dominant Polycystic Kidney Disease (15 exomes and 3 panel sequencing datasets). Compared to Bowtie2, BWA-MEM2 and minimap2, ParaDISM reduced the number of misalignment artifacts, thus increasing specificity and precision of subsequent single nucleotide variant calls. Conclusions ParaDISM refines BAM files to improve the precision of read placement and downstream analyses in highly homologous sequences. The pipeline can be used as a precision-focused add-on to standard bioinformatic pipelines before the variant calling step to provide a stronger level of evidence for variants called in regions of high homology. This approach is useful for applications where false positive variant calls have more severe consequences than false negative ones, such as identifying causal mutations for genetic disorders. In applications where experimental wet-lab confirmation of shortlisted variants is necessary, ParaDISM can be used to prioritise variants with strong support to reduce cost and effort spent on attempting to confirm false positive calls. ParaDISM can be used alongside the results of standard mappers to cross-check results and identify unreliable read placement and variants potentially caused by misalignment artifacts. The pipeline is open source and available under the MIT license at github.com/BioGeMT/ParaDISM.