The scrapheap challenge: extracting relevant data from unmapped whole genome sequencing reads.. The scrapheap challenge

Unmapped sequencing reads are typically ignored while they contain biologically relevant information. We systematically analyzed unmapped reads from whole genome sequencing of 33 inbred rat strains. Reads were filtered for low quality, reference omissions and contamination to enrich for biologically relevant sequences. Similarity-based clustering of filtered reads revealed clustering similar to previously reported phylogenetic trees. Our results demonstrate that 20% of unmapped reads do harbor relevant information that can be used to improve biological knowledge, including potential genotype-phenotype relationships. Analysis pipelines would benefit from incorporating the described methods and reference genomes would benefit from inclusion of the obtained genomic segments.