BECAS
FEMENIAS martin Miguel
congresos y reuniones científicas
Título:
The use of Selective Alignment to explore transposons derived from RNAseq data.
Autor/es:
FEMENIAS, MARTIN MIGUEL; SANTOS, JUAN CARLOS; SITES, JACK W.; AVILA, LUCIANO JAVIER; MORANDO, MARIANA
Lugar:
Uppsala
Reunión:
Seminario; 5th Uppsala Transposon Symposium; 2021
Institución organizadora:
Uppsala University, University of East Anglia, Karolinska Institute
Resumen:
Given the repetitive nature of transposable elements (TEs),identification and quantification of TEs pose a challenge, and theyare usually missed in the RNAseq data analysis workflows. TheTEtranscripts algorithm performs a read assignment phase followed bythe maximum likelihood estimate of multiple-reads assignments usingthe expectation-maximization algorithm. Although this procedureprovides reliable estimates, the computational cost is high.Alternatively, tools with a lightweight mapping approach considerablyreduce the computational cost at the expense of mapping accuracy.Yet, neither of these tools consider the origin of the reads, andreads from non-exonic regions (e.g., non-removed introns or UTRregions) might introduce additional bias. Here we propose a selectivealignment procedure (SA) for the exploration of TEs fromtranscriptomic data. The SA increases the mapping sensitivity andperforms an alignment scoring phase that allows spurious mappings tobe filtered out. The combination of SA with decoy sequences avoidsfalse assignments and allows to consider the origin of the reads,avoiding the bias introduced by reads derived from repetitive regionsco-expressed with genes. For this purpose, we present an R package(ExplorATE) for the exploration of TEs in RNAseq data with or withoutreference genome and we show that this procedure achieves a precisioncomparable to TEtranscripts but with a computational cost of 5 to 32times lower.p { margin-bottom: 0.1in; direction: ltr; line-height: 115%; text-align: left; orphans: 2; widows: 2; background: transparent }