[khmer] Khmer for

Richard Smith rds45 at cam.ac.uk
Sat Nov 30 11:19:32 PST 2013


You could use either the kmer counts or estimated expression counts to try to find the cell type of origin as follows:

Take the publicly available chicken cell type expression datasets and process them all the same. Make a matrix with each gene/transcript/kmer as a column and each sample as a row. Cells contain the counts. Then you've reduced the problem to a clustering or classification problem, and can use, for example, hierarchical clustering to which cell type your sample is closest to.

Sent from my iPhone

> On 30 Nov 2013, at 18:03, Tamer Mansour <drtamermansour at gmail.com> wrote:
> 
> Dear Colleagues,
> I am working on sequencing project for tumor tissue in chicken. The cell of origin is not known. My idea is to use the Khmer package to compare the unique khmers and their abundance in our tumor tissue to the publicly available expression seq to identify the closest one. Any suggestion about the idea? Likit said he has script that do the comparison already, hopefully he can find it and share it with us.
> 
> Thank you
> 
> Regards 
> 
> 
> Tamer     
> _______________________________________________
> khmer mailing list
> khmer at lists.idyll.org
> http://lists.idyll.org/listinfo/khmer




More information about the khmer mailing list