Mapability

Mapability, often spelled mappability, refers to the ability to uniquely assign a sequencing read to a location in a reference genome. It is influenced by the underlying sequence repetitiveness, the chosen read length, and the error profile of the sequencing technology. In practice, mapability is represented as a per-base score or as a regional metric indicating how reliably reads from that region can be mapped.

Mapability is typically estimated by examining all possible reads of a given length from the genome and determining how many locations those reads map to exactly or within a defined number of mismatches. A base supported by a unique L-mer is highly mappable; if the L-mer occurs in multiple locations, it is less mappable. Tools and data tracks provide per-base scores or region summaries for specific read lengths, such as 50, 76, or 100 bases, and may include allowances for mismatches or sequencing errors.

Applications of mapability include informing variant calling and copy-number analysis, interpreting read-depth signals, guiding the design of sequencing experiments and capture probes, and annotating genome regions by their reliability for mapping. In analysis workflows, low-mapability regions are often masked or down-weighted to reduce spurious signals arising from ambiguous read alignment.

Limitations arise from dependence on the reference genome version, alignment algorithms and parameters, and the error characteristics of sequencing data. Mapability predictions assume ideal conditions and may not capture all practical biases of real data, especially in highly repetitive or structurally variable regions. Consequently, observed mapping performance can differ from predictions, underscoring the need to consider mapability alongside other quality metrics in genomic analyses.