HELP! How to interpret this eland output

[skblazer]

>ILLUMINA-053F9F_0003:8:1:1051:4272#0/1 GTTTTAATCGATAACGGGAGATGTGGTGTTCGGGAGGACCCGCTTCACACTAATAAAGCGTAACCCTACAATCCAAACAATCTCG 2:0:0 silkworm2.fa/scaffold1100:8972F37TT6TCTCG1CGCG1C1AC1CCTGTCCA1G1T1GTCTTGAACG1,silkworm2.fa/scaffold1143:2756F37TT6TCTCG1CGCG1C1AC1CCTGTCCA1G1T1GTCTTGAACG1


I want to know that

1) what does the last string after '2756F', that is '7TT6TCTCG1CGCG1C1AC1CCTGTCCA1G1T1GTCTTGAACG1', mean?

2) Actually only the first 37bp of this read (85bp) can align the reference by my blast verification, so why eland report it as mapped with less than 2 errors?

Query: 1 gttttaatcgataacgggagatgtggtgttcgggagg 37
|||||||||||||||||||||||||||||||||||||
Sbjct: 2756 gttttaatcgataacgggagatgtggtgttcgggagg 2792

[skblazer]

Please help!

[mattanswers]

ELAND creates a lot of files, so first it is helpful to identify which file; the s_*_export.txt, s_*_sorted.txt, s_*_sequence.txt, etc. that you are referring too.
You can also look here:
http://seqanswers.com/forums/showthread.php?t=330
I would guess that you matched to two sequences in your reference genome, silkworm2.fa. The first, scaffold1100:8972, and the second, scaffold1143:2756. The alpha-numeric string after those two numbers, F37TT...CG1 are both the same and I would say are quality scores for your sequence. Each letter or number symbolizes a probability that the corresponding base in your sequence in correct.

Could your reference sequences that were matched be only 50 bps ?

The two errors may be only for the first 28 bases, but that I am not sure about.

[swbarnes2]

>ILLUMINA-053F9F_0003:8:1:1051:4272#0/1
2) Actually only the first 37bp of this read (85bp) can align the reference by my blast verification, so why eland report it as mapped with less than 2 errors?

Once upon a time, Eland only used the first 36 letters for aligning, and just tacked the rest on as an afterthought. So Eland still works that way, then that's expected behavior.

I'd switch to a better aligner. Bowtie, or SOAP2, or bwa...anything that uses a Burrows-Wheeler Transform algorithm, because they are faster. Bowtie won't do indels, but bwa will.