Tophat mapping Results

[figo1019]

Hi All

I used tophat to map my 100 million base paired illumina reads after trimming with average length of 86 bp. In the bam output file generated by tophat I only find around 55 million read mapped and around 42 million reads unmapped. Can any one tell me that is it a feasible number in case of mammals.
Also what can be the reason on so much of unmapped reads since I have already cleaned the data with q value 30 and removed the adapters and primers.
In this paper http://www.plosone.org/article/info%...l.pone.0046415 I see that the comparison of topaht with GSNAP where tophat has very low mapping reads

Regards

[chadn737]

1) What are the commands you ran Tophat with?
2) What is your reference? I don't think one can generalize about mammals or plants or anything like that.
3) What I always like to do when I have a lot of unmmapped reads is take a random selection of those reads and just Blast them and see what pops up. For example, I once had some reads with a lot of adapter sequence in the 3' end. I found that out when my Blast results showed that only part of the read matched my reference. I have also worked with libraries with a lot of reads from the mitochondria, for which I did not have a completed genome for. I know this only because I blasted those reads to see what they might be.