Tophat and Bowtie results

[baohua100]

i have reads.fq (5 000 000 reads RNA-seq) genome.fa

when i use bowtie align the reads, it aligns 4 200 000 reads.

but when i use tophat align the reads, the output file accepted_hits.sam only contains 1 900 000 reads. junctions.bed file is blank.

[chengeng]

Hi,
Could you tell me the parameters you set?

[baohua100]

default parameter


tophat ref.fasta reads.fq

[sorrychen]

I also have similar question of inconsistent number of mapping, probably because the parameter setting. Any suggestion about setting the parameter in bowtie/tophat so I can take advantage of the long reads while allowing more mismatches. The /tmp/segments_juncs.fa only have 45bp. How can I use longer reads to map to the junction?

[chengeng]

Hello,
You can send E-mail to Cole Trapnell([email protected]). He is the author of TopHat and Bowtie.

All the best,

Geng

[chengeng]

Hello,
You can email Cole Trapnell([email protected]).He is the author of TopHat and Bowtie.

All the best,

Geng

[sorrychen]

Thanks, and sorry about the late response. I was told (and confirmed by the options) that Tophat cut long reads into smaller pieces to find junction. In this way, 45 bp junctions will be enough. I think the design is try to avoid there are super small exon that is smaller than 45 bp. If so, a 100 bp reads may mapped to 3 exons.