SplitRead

Splitread, or split-read mapping, is a genomic sequencing concept used to detect structural variation by identifying sequencing reads that align to two separate genomic loci. This phenomenon occurs when a read spans a breakpoint created by insertions, deletions, inversions, or translocations, causing part of the read to map to one region and the remainder to another. Modern aligners can output split alignments or soft-clipped segments, enabling downstream tools to infer precise breakpoint locations.

In practice, split-read analysis combines split-read evidence with other signals, such as discordant paired-end mappings and read-depth changes, to call structural variants. After mapping reads to a reference genome, specialized software clusters split-read alignments that support the same breakpoint, refines breakpoints to base-pair resolution, and produces a list of candidate structural variants with supporting evidence.

Splitread is particularly effective for identifying small to medium-sized variants and for resolving breakpoints with high precision, which is valuable in cancer genomics and germline SV discovery. It complements methods based on paired-end mapping and read depth, which may have lower resolution or miss complex rearrangements.

Limitations include dependence on read length and genomic mappability: repetitive or low-complexity regions can produce ambiguous alignments, and reliance on sufficient coverage means some events may be missed in shallow datasets. Although powerful, split-read results typically require validation and are often integrated with other SV-detection signals to reduce false positives.