Pl see the link for answer

https://link.springer.com/article/10.1186/s13059-016-1103-0

Oi Danillo,

How many reads do you have, how big is your reference - is it a Eukaryot?

The reads can be of quite low quality, but I would have expected that blast should map something at least against the transcripts (CDS), if you have enough reads. Maybe lower the word size for the blast -W parameter.

Last, it could be also contamination, so blast the reads against nt database on NCBI.

I would try that first before a de novo assembly. (Canu is a good option but not sure if you have enough coverage and how it perform and the uneven nature of the RNA, you won't get all the genes.)

Hi Danillo

since the Minion can give data with a high error rate, do you have an idea on the quality of the samples you're working on? maybe the addition of low quality samples and high error rates would only give you the impossibility to align, whatever the tool you use to do that.

fred