Tweet
Hi All,
The goal is differential expression analysis of a species of fly and I'd appreciate some opinions on multiplexing a large number of samples (60+) and sequencing several technical replicates of the pool to achieve the desired read depth.
I expect the low depth per sample (~4M) will cause sampling variance and therefore technical variation among runs to be large. Is it acceptable to simply combine the data from each run and ignore the fact that different runs were used (HiSeq2500 or NextSeq500)? Alternatively, despite having to account for lane effects, would a complete block design using 12 sample multiplexes (~20M reads per sample) be a better option?
Thanks.
The goal is differential expression analysis of a species of fly and I'd appreciate some opinions on multiplexing a large number of samples (60+) and sequencing several technical replicates of the pool to achieve the desired read depth.
I expect the low depth per sample (~4M) will cause sampling variance and therefore technical variation among runs to be large. Is it acceptable to simply combine the data from each run and ignore the fact that different runs were used (HiSeq2500 or NextSeq500)? Alternatively, despite having to account for lane effects, would a complete block design using 12 sample multiplexes (~20M reads per sample) be a better option?
Thanks.