Mapping reads from different genus

[Break]

Hi everyone!

There are RNAseq reads for a mammalian specie with unsequenced genome. The most close mals genome that have been sequenced is related to the different genus. Did anyone try to map reads on so phylogenetically distant genomes (with bowtie/tophat)? Is there any sense to do this or it is better the de novo assemblie in this situation ? Thank you.

[Break]

PS: such a mapping is necessary for expression value calculations....

[ssing]

I have tried mapping to a genome from an organism about 15-20 mya from the organism of interest. I found that I only had a slight decrease in mapping efficiency (90 -> 80%) but that more divergent stuff didn't map well at all. I imagine you could try doing this iteratively, and map more stringently and then loosen the parameters and map again. FWIW, I ended up assembling de novo and then mapping.

Good luck!

[Break]

Thanks for the reply. De novo is a possibility. However, there is another question. I have several tissues. Is it better to pool all samples and map than reads to the reference contigs? Or it is better to do assembly for each tissue in separate because of presence tissue-specific isoforms?