a Question about Duplicated Genes

[aquleaf]

Hi all.

I have a question about duplicated genes which have several copies in the genome such as Ccl21c. Tophat was used to map the RNA-Seq reads back to the genome and HTSeq was used to count the reads map to each gene. All of the reads stem from the duplicated genes will be get rid of by HTSeq because they are aligned to multiple places. Is there anyway to quantify these genes? Or could I mask the genomic regions of duplicated genes before running Tophat?

Any suggestion will be much appreciated.

Best

[swbarnes2]

Don't mask the genome.

The last thing you want is reads being forced to align to the wrong place, because you masked away the right place.

If the genes are true duplicates, it's going to be pretty impossible to separate out what reads cam from where.

[aquleaf]

Thanks very much for your reply. We don't want to distinguish which locus these reads come from for a gene with several copies in the genome, we wish to count all the reads uniquely mapped to this gene. For example, if a gene has 4 copies in the genome, we want to count the number of reads mapped to those regions which were not mapped to other genes. Is there any way to achieve that?

One more request, could anyone recommend a tool to modify GTF file used in the RNA-Seq analysis? We used the reference GTF file from UCSC and found lots of genes seem identical, such as Gm14430, Gm4724 and Gm14434. We wish to merge such genes into a single gene.

Thanks a bunch!

Best

Don't mask the genome.

The last thing you want is reads being forced to align to the wrong place, because you masked away the right place.

If the genes are true duplicates, it's going to be pretty impossible to separate out what reads cam from where.

[JackieBadger]

http://bioinformatics.oxfordjournals.../2711.abstract

http://genome.cshlp.org/content/20/11/1613.abstract

[liux]

We mask all but one copy if the duplicates are exactly the same (~50% of genes) or less than 1% (other 40%); and flag the rest.

It is probably OK if just looking at the gene expression. recently we start to integrate ChIPseq data with mRNAseq data. I can see this approach will cause problems there.

[aquleaf]

We mask all but one copy if the duplicates are exactly the same (~50% of genes) or less than 1% (other 40%); and flag the rest.

It is probably OK if just looking at the gene expression. recently we start to integrate ChIPseq data with mRNAseq data. I can see this approach will cause problems there.

Thanks very much for your reply. How do you mask them? Is there any software to mask them?

We will also combine the ChIP-Seq data with mRNA-Seq data later on. What's the problem if the duplicates were masked?

Thanks very much!