multimapped

Multimapped refers to sequencing reads that align to multiple locations in a reference genome or transcriptome with similar alignment scores. This situation commonly arises with short reads and repetitive sequences, such as transposable elements, segmental duplications, and highly conserved gene families, where several genomic locations share near-identical sequences.

Multimapped reads pose challenges for downstream analyses because assigning a read to a single genomic location can introduce bias or uncertainty. In RNA-seq, for example, reads originating from paralogous genes or repetitive exons may map to several loci, complicating gene or transcript quantification. In ChIP-seq or DNA sequencing, multimapping can affect peak detection and copy-number estimation.

Strategies to handle multimapped reads vary. Some analyses discard multimapped reads altogether. Others retain them but rely on mapping quality scores or report all possible locations. Probabilistic approaches distribute read counts across potential locations according to estimated abundance (for example, expectation-maximization methods used in transcript quantification tools). Other methods randomly assign one of the possible locations, or use genomic mappability information to down-weight or mask low-mappability regions.

Tools and parameters used to manage multimapping are common in alignment and quantification workflows. Aligners may limit the number of reported locations or provide options to output all alignments with associated scores. Transcriptome-aware quantifiers often incorporate multimapping handling to improve estimate accuracy.

In summary, multimapped reads reflect intrinsic genomic repetitiveness and read length limitations, and their handling is a key consideration in ensuring accurate interpretation of sequencing data. A contrasting term is uniquely mapped reads, which align to a single location with high confidence.