Right way to convert splice-aligned Solexa reads (s_?_export.txt) to SAM format

[jorgebm]

Hello,


I would ask you about the right way to convert splice-aligned Solexa reads (s_?_export.txt) to SAM format.

Some time ago I've performed a "eland_rna" analysis using GERALD.pl v.1.15 (CASAVA 1.6). Now I exported s_?_export.txt file to SAM format using

"illumina_export2sam.pl" script (bundled with CASAVA 1.7).

Focus on splice-aligned reads I found weakness in this format conversion script. For instance,

this s_?_export.txt read line:

HWUSI-EAS1597 1 7 1 1 1467 0 1 ATGGAAGCTCTGCGGTNATACAACCAGGAGCACTC aa^`]aa`a`a_`]`ZB`___[a_`XZWU[UWTUZ splicesites34.fa RIC8A_34_34_chr11.fa_198938_199271 5 F 16C18 105 Y


is converted to the following SAM line

HWUSI-EAS1597_1:7:1:1:1467 0 splicesites34.fa/RIC8A_34_34_chr11.fa_198938_199271 5 105 35M * 0 0 ATGGAAGCTCTGCGGTNATACAACCAGGAGCACTC BB?A>BBABAB@A>A;#A@@@<B@A9;86<6856; XD:Z:16C18 SM:i:105


This splice is (from refFlat)

chr11 hg18_refFlat exon 198530 198938 0.000000 + . ID=exon:RIC8A:1;Parent=RIC8A;
chr11 hg18_refFlat exon 199271 199318 0.000000 + . ID=exon:RIC8A:2;Parent=RIC8A;

And in Illumina splicesites34.fa format

>RIC8A_34_34_chr11.fa_198938_199271
GATTATGGAAGCTCTGCGGTCATACAACCAGGAGCACTCCCAGAGCTTCACGTTTGATGATGCCCAAC

So, the alignment is

------ATGGAAGCTCTGCGGTNATACAACCAGGAGCACTC (read)

GATTATGGAAGCTCTGCGGTNATACAACCAGGAG (last 35b of RIC8A:1)
------------------------------------------------------CACTCCCAGAGCTTCACGTTTGATGATGCCCAAC (first 35 bases of RIC8A:2)



Note that field 4 of SAM line (POS), that is formally defined as "1-based leftmost POSition/coordinate of clipped sequence", is set to 5. That's the offset against splicesites34.fa reference file and not to the genome reference files (chr11.fa).

I think it's wrong. it isn't?. I mean this is not a full compliance format conversion.