On Wed, Feb 18, 2009 at 1:29 PM, Wolfgang Raffelsberger <
[email protected]> wrote:
> Dear List,
>
> here I have a more general question :
> Has anyone compared (directly?) Affy 500kSNP arrays with deep sequencing
> (here I'm rather thinking of Solexa/Illumina) ?
> How much sequencing coverage should allow a 'decent' SNP calling ? (Maybe I
> need to add that heterozygous SNPs should be detected, too)
> At which level (of coverage) would you expect to get equivalent or better
> results as Affy SNP arrays ? (Of course I know that a few (specific) SNPs
> may be lost due to the random distribution of the reads...)
>
You could look at the 1000 genomes pilot data available at SRA. They have
sequenced some of the individuals (I think a trio or two) from hapmap at 10x
coverage (approximately), if I recall. You could look for mendelian
inheritance errors and for concordance with the SNP data.
>
> Or a variant of the question, if one sequenced a given zone of high
> interest (possibly with multiplexing) what would be current advice in terms
> of size of the zone (and number of multiplexed patients) to get suffient
> coverage to detect heterozygous SNPs.
>
These questions depend on the error characteristics of the sequencing
platform. However, I think a number of 20-30x coverage is probably not too
far off, at least for solexa. And, of course, the uniformity of any
selection technology has to be taken into account when figuring out the
actual coverage obtained.
Hope that helps.
Sean
[[alternative HTML version deleted]]
_______________________________________________
Bioc-sig-sequencing mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/bioc-sig-sequencing
