assembly unmapped reads from tophat

[leifive]

Hi,
I have used Tophat/Cufflink to map and assembly my RNA-seq pair-end data(the insert size is 200bp while each read has 90bp length) with a reference genome. Now I am thinking how to de novo assembly the unmapped reads so that i can find some noval transcripts out of the reference genome. My tophat version is 2.0.8, so i got a output file named unmapped.bam. I am running bam2fastx with -o and -P options but running into errors. It generates a fastq without -P option, but i am not sure whether i can use this single fastq file to do a de novo assembly. Could anyone help me?

[yzzhang]

I used bam2fastq tools, it can detect whether the reads in the bam file are pair end, and generate reads1 and reads2 file automatically if it is true.

[leifive]

I used bam2fastq tools, it can detect whether the reads in the bam file are pair end, and generate reads1 and reads2 file automatically if it is true.

Thanks, yzzhang. I have installed bam2fastq and running on my 500MB size unmapped.bam file. But converting it to paired end fastq files is taking much longer than the original tophat run which was only last 2 hours. One night has passed, and now the pair-end file size only reach 15MB. Is there someting wrong? or all i need is just waiting?

[leifive]

Oh, It's my fault. I forget to run bam2fastq with -readname option. It's OK now. Thanks again.