What are key factors to consider while planning pooled DNA sequencing

[San220]

Hi,

I am planning to sequence two pooled DNA samples. I was wondering if there is:

1) Minimum number of individuals that should be pooled to make up each sample. More specifically, would pooling of 2-3 individuals per sample be enough to detect sufficient number of variants to distinguish it from sequencing errors?
2) The methods that can be conducted to eliminate false positive SNPs from the generated data.

Thanks and best regards!

[nucacidhunter]

If you post more info on your planned work such as sample numbers, genome size and aims your question will be more clear.

[sarvidsson]

Is there a specific reason why you want to pool the samples, and not make separate libraries for each? Except for saving money on library construction, there is no benefit in pooling, rather the opposite.

1) The more individuals you pool, the more difficult it will be to distinguish variants from sequencing errors, as you are diluting the signal for rare variant alleles.
2) The simplest method to reduce the number of false positives is brute force - increase the sequencing depth per sample. There are several bioinformatic methods implemented in current variant callers (e.g. GATK) to account for typical sequencing biases leading to false positive calls.

If your main interest is comparing different conditions (represented by large groups of individuals, e.g. disease vs control groups), you would typically need more individuals (think of 20-30 per group rather than 2-3) unless there are strong reasons to expect single variants to explain the condition (which is almost never the case). Still single libraries would give you much more useful data and more statistical power than pooled libraries.

[San220]

You're right about the objective of the study Sarvidsson, which is to conduct variant analysis between fast (plant root) growth-rate vs slow growth-rate groups with 60-70x coverage for each of the two libraries.
I was hoping to find more than 2-3 individuals per group,unfortunately that didn't happen and I ended up with a greatly reduced pool-size. Financially, I can only sequence two libraries which means I either scrap this idea or go ahead with sequencing two, small pool-sized groups.
What do you guys think? Thanks