eland realign file

[qiudao]

we have recently received ChIP-seq results with using illumina sequencing. We got a realign file. there is a part of realign_readme, I can not really understand.

"Output description: The first 12 lines are the header that provide information pertaining to the Chastity filtering.
Following the header lines, either three or seven columns are generated depening on whether the sequences were aligned to a string of A's or a reference genome respectively......."

What does this "the sequences were aligned to a string of A's" mean?

How was the realign file generated? Thanks.

q

[Chipper]

Hi,

as I understand it the realign files contain only those sequences that pass the probability criteria for the first 12 bases (p >0.6). It can be aligned either to genomic sequences, then you get additional colomns with alignment info or by aligning to a stretch of A:s, then you do not get the positions.

For ChIP-seq you may want to use a less strict criteria on the sequences though, as many that do not pass the filtering still get mapped to the correct genomic location.

[bioinfosm]

What about RNAseq, do you think using 'passing filter' reads is too conservative?
We feel that checking first 12 bases for anything p<0.6 is too strict overall.

[Chipper]

I have not done RNAseq on the Illumina, but you could check what the distributions look like, i.e. what % are in exons in those that pass the filtering and in those that does not. But then since many exons are similar, e.g in gene families, it may be more important to have a low error rate?