I'm doing a FACS based CRISPR screen. Due to the large physical size of the cells, I am unable to sort extremely large numbers of cells at once.
To get an adequate number of cells for good coverage, I'm doing multiple sorts/experiments that I would like to pool. However, I'm worried about biasing the library if one experiment is under/over represented in the pooled product.
So far, I've got:
Exp1: One sample was sorted into population A and B:
Population A: ~16M cells ->150 ug of gDNA
Population B : ~16M cells -> 110 ug of gDNA
Exp2 : One sample was sorted into population A and B:
Population A: ~15M cells -> 110 ug of gDNA
Population B : ~15M cells -> 100 ug of gDNA
1. I'll do a few more passes for more cells.
2. Then I could pool all the "A" together for adequate coverage.
3. Then I'll sequence and compare "A" vs "B"
Exp1's "A" is over-represented in the pooled "A"
Exp2's "B" is over-represented in the pooled "B"
I don't think that would be very good. Exp1 and Exp2 probably had different initial guide biases in their pre-sorted populations, which would confound my results.
What would be the best way to pool the samples? Or should I perhaps never pool them, and instead use different Illumina adaptors for both Exp1 and Exp2 (and the remaining experiments)?