Creating a SNP library

[bluefin]

Hi we are trying to create a SNP library for roughly 250 SNPs across the human genome. I am fairly new to SNP detection with Illumina HiSeq so could anyone direct me to protocols or studies for methods that are most cost effective.

[Number6]

We've been running a number of targeted SNP genotyping projects, which are amplicon libraries based on Illumina's two-step PCR amplicon approach. Illumina has an app note. It uses Nextera tails on the locus-specific primers, followed by the Nextera dual-indexed "universal" primers.

With this strategy, we've been running the first PCRs with ~24 multiplexed primer pairs, and pooling these from 384 individuals. To get your 250 loci, you could simply run ~10 multiplex reactions for each individual (or run a higher multiplex.....haven't tried it, but it might work). You'd then be able to pool all 250 amplicons from a single individual before indexing, and pool everything together for the sequencing run. We've been running the MiSeq, but with your 250 loci x multiple individuals, a HiSeq will likely be better.

Each read is tagged by the locus-specific primers, which you read through before you get to the SNP site and with the large number of amplicons, complexity is not an issue. Your amplicons can be short, which helps with amplification efficiency but amplification efficiency is not much of an issue if you're simply looking to call homozygous/heterozygous. Who cares if the read numbers vary 100-fold between loci/individuals, as long as you get hundreds of reads from each locus? (back of the envelope calculation, for 250 loci x 96 individuals, with 180 million reads, you'd get on average 7500 reads for each).