Jitterbug: somatic and germline transposon insertion detection at single-nucleotide resolution

Hénaff, Elizabeth; Zapata Ortiz, Luis, 1985-; Casacuberta, Josep M.; Ossowski, Stephan

doi:http://dx.doi.org/10.1186/s12864-015-1975-5

Jitterbug: somatic and germline transposon insertion detection at single-nucleotide resolution

License

Citation

Hénaff E, Zapata L, Casacuberta JM, Ossowski S. Jitterbug: somatic and germline transposon insertion detection at single-nucleotide resolution. BMC Genomics. 2015;16(1):768. DOI: 10.1186/s12864-015-1975-5

Abstract

Background: Transposable elements are major players in genome evolution. Transposon insertion polymorphisms can/ntranslate into phenotypic differences in plants and animals and are linked to different diseases including human cancer, making their characterization highly relevant to the study of genome evolution and genetic diseases./nResults: Here we present Jitterbug, a novel tool that identifies transposable element insertion sites at single-nucleotide resolution based on the pairedend mapping and clipped-read signatures produced by NGS alignments. Jitterbug can be easily integrated into existing NGS analysis pipelines, using the standard BAM format produced by frequently applied alignment tools (e.g. bwa, bowtie2), with no need to realign reads to a set of consensus transposon sequences. Jitterbug is highly sensitive and able to recall transposon insertions with a very high specificity, as demonstrated by benchmarks in the human and Arabidopsis genomes, and validation using long PacBio reads. In addition, Jitterbug estimates the zygosity of transposon insertions with high accuracy and can also identify somatic insertions. Conclusions: We demonstrate that Jitterbug can identify mosaic somatic transposon movement using sequenced tumor-normal sample pairs and allows for estimating the cancer cell fraction of clones containing a somatic TE insertion. We suggest that the independent methods we use to evaluate performance are a step towards creating a gold standard dataset for benchmarking structural variant prediction tools.