Detecting differentially expressed genes using aligner outputs

[questioner]

Hi all,

I am new to the forum, and I am working with a Illumina GA II data. I have done sequence alignment using Bowtie. My goals is to detect differentially expressed genes over two populations just like microarray analysis. Can someone recommend an easy-to-use software to perform this task?

I have tried many, and the pile-up functionality does not solve my problem. What I'd like to have is to retrieve counts for each short read with genome annotation such as gene name etc.

Thanks very much!

Questioner

[swbarnes2]

Parse the Bowtie output directly. If you aligned straight to named transcripts, each line will have which transcript the read aligned to. It will also say whether the read was uniquely aligned or not. So just count, and correct between your samples based on total number of reads, or reads aligning to housekeeping genes.

If you aligned to whole genome, you'll have to work out from the position which read aligned to which transcript, but that's doable.

[krobison]

Have you looked at TopHat or ERANGE? ERANGE in particular would seem to fit your request. (I haven't tried either; am new to this too)

[questioner]

Thank you all. Now I am parsing Bowtie output and confused about the output format of Bowtie. See my another post please.

[zee]

ERANGE will parse bowtie output and create read datasets for you, therefore no need to write your own bowtie parser.
Read the ERANGE documentation, it's quite easy to follow.
The Bowtie manual also gives a full description of all the output fields.

[questioner]

Thanks for the response. I will try ERANGE.

I read Bowtie manual carefully and it describes 8 fields in the output table, but I got 15 fields in my run. Don't know why?

[waseem]

-sam format gives you more fields (including optional fields). using --sam-nosq, will give you 8 fields, the fields you don't want can be suppressed with -suppress 1,2,3 etc.