unmapped reads - bowtie

[csoong]

there is an option to dump the unaligned reads in bowtie.
you may check to see whether the quality of those reads are bad or not.
also, if you are aligning mRNA, you shouldnt be aligning against the reference genome. since a transcript would be genome sequence minus introns and stuff like that.
since bowtie cant align gaps, all those mRNA reads will show as unmapped.

[frymor]

Hi everybody,

I am running bowtie to analyse my paired-end Illumina data sets. I have two files one from wild type and one from a mutation we are working with.

to run bowtie i used this command:
bowtie -a --best --strata -m 10 -n 2 -l 22 -q --un dilptotal.unmapped -t -p 6 --chunkmbs 256 --max dilptotal.maxHits -S d_melanogaster_fb5_32 -1 dilp_s2_1_sequence.fq -2 dilp_s2_2_sequence.fq dilptotal.sam

and to run it with the option of trimmed ends I used this command:
bowtie -a --best --strata -m 10 -n 2 -l 22 -q --un dilptotal_trimmed.unmapped -t -p 6 -3 11 -5 11 --chunkmbs 256 --max dilptotal_trimmed.maxHits -S d_melanogaster_fb5_32 -1 dilp_s2_1_sequence.fq -2 dilp_s2_2_sequence.fq dilptotal_trimmed.sam

Beside the fact that only later, I read that the options 'best' and 'strata' are not working with paired-ends , I run it as it is and got the following results:
for the option w.o trimming:
Read number 24781053 (57.81%)
Unmapped 17831504 (41.60%)
Max. hit 255358 (0.60%)
for the trimmed command:
read number 28,575,776 (66.66%)
Unmapped 14,003,690 (32.67%)
Max. hit 288,449 (0.67%)

As one can see here I have over two third of my data mapped when I am trimming both ends (while only 57% with the complete reads). If I understand it correctly I can assume that the ends are adapter residues which were needed to be cut out (I run quality control using fastqc software).

My question here regards the unmapped reads. I can see that there are almost a third of my data which wasn't mapped to the genome. Is there a way of telling why this reads where not mapped.
These are the reason i could think of for a read to not be mapped:
1. quality of read is too bad
2. repetitive elements - more than 10 possible hits (with this preferences).
3. reads were not found on the reference genome
4. ???
Can you please tell me what other possible reasons there can be for not being able to map a read?

Is there a way of telling what reads are falling into which group?

What abut the reads which were mapped. As I am working with RNA-Seq data I am interested to know if I have some rRNA residues in my reads.
Is there a way of finding out how much of this two third mapped reads are really mapped to mRNA and not to some rRNA (which was left over after cleaning/purification of the mRNA) or even to some intergenic region or interons?

Thanks for any help
Assa