Closed Hoohm closed 7 years ago
Hi Patrick,
all the BLAST alignments with 100% identity are very short, <22b. I also mapped it with BLAT and it gives only one short hit: 22 2 23 49 100.0% 11 - 4703206 4703227 22 If you want to output such short alignments, you would need to reduce minimum score and mapped length requirements, e.g.: --outFilterScoreMinOverLread 0.3 --outFilterMatchNminOverLread 0
Cheers Alex
Hi,
thanks for the help. I tried the changes and I actually got ~10-20% more uniquely mapped reads. no more too short and the rest moved to multi mapping (which seems reasonable).
Thanks a lot!
Applying --outFilterScoreMinOverLread 0.3 --outFilterMatchNminOverLread 0
made the too short reads map to multiple loci instead
Hello,
I've been having some problems with a few runs of our single cell RNAseq data. We get a high percentage of unmapped reads. I've checked a couple of potential issues but it doesn't seem to solve them. 1) I've change the --sjdbOverhang to 48 (our reads are 49bp) --> no change 2) I'm using now --twopassMode Basic I might be missing some options here. There is an example of a sequence that I found in the unaligned tagged uT:A:1 CTTTAATTTTATTTGCTTATGTTACCATGCACAAAGGTTTCAGTTACTA
This sequence does map to mouse, here is a blast result:
Here is the Log.out
Don't really know what else I should try. The sequence maps on Zfp276 in the mouse Genome. Do you have any idea which parameters I should tweak to get those mappings? Thanks