low percentage of reads mapped

[rahilsethi]

Hello,

I was analyzing the SOLiD SAGE data of human sequence using SOLiD SAGE Analysis tool. I performed Mapping using 27 bp length with 1 mismatch. The reference was the complete set of human mRNA sequence from Refseq db. I then calculated % of reads that mapped to reference using the results file that gives the list of tags and their corresponding read files. I ran the analysis for 4 SAGE data and I got the following percentage:
SAGE A : 15 % apporx.; SAGE B 16% approx.; SAGE C 17% and SAGE D 20 %

What can be the reason for such low percentage of reads mapping to human mRNA reference? Is it a general result for most of the SOLiD SAGE experiments?

Thanks

[NextGenSeq]

BLAT some of the unmapped reads to see what they are. If they are actually real mRNA, its your analysis. If they are genomic DNA or something else, it's your library.

[rahilsethi]

What it has to do with my analysis? The result is straight from the software SOLiD SAGE Analysis tool. In the result.tab file it produces read ids for the tags are mentioned. I counted the unique set from those read ids and divided it by the unique set of read ids from the read file to get the percentage of reads mapped to human mRNA obtained from Refseq database.
I will still BLAT some of them and see where they are mapping
If they are not mapping to mRNA then something should be with the reads generated by SOLiD SAGE run

[pmiguel]

Hi Rahlisethi,

NextGenSeq is giving you a "sanity check" to help with your troubleshooting. The SAGE might be working fine but your RNA might contain lots of transcripts from repetitive elements, for example. Then their position is not uniquely mappable in the genome. Or you might get hits to E. coli, or some other unexpected species, and suspect contamination from some source.

--
Phillip