Open markusglass opened 5 years ago
Why don't you map them with to the transcriptome with minimap2 and use salmon to quantify the alignments? Seem like with such long reads you might also want to assemble them first and add those transcript models. Your UTRs might be driving a low mapping rate, since for genomic alignment you'll map the whole UTR, the length of which is usually not captured well in canonical transcript models.
Hi,
I wanted to quantify reads sequenced with a Oxford Nanopore MinION using salmon 0.12.0. With activated --validateMappings option mapping rates were below 1% (mapping them to the human genome using minimap2 gave me ~95% mapping rates). When I didn't set the --validateMappings flag I got ~8% and with activated --validateMappings and very low --minScoreFraction (0.01) I could achieve at most 18% mapping rates. --fldMean was set to 800 in each case (according to the mean read lenghts reported by NanoPlot). Is there a possibility to improve the mapping rates on these noisy long reads?
Regards, Markus