Sequencing depth needed

[sequencenewbie]

Hi All,

I'm interested in doing some transcriptome profiling to identify differentially expressed genes between two populations.

how much sequencing depth do I need to do a meaningful study? In the literature I've seen reports of 100 million reads on SOLID, 40-50 million on illumina, and down to 10 million 25bp reads on illumina?

Hoping to get some advice.

Thanks in advance

[knightfeng]

I think the most intuitive way is build relationship between sequencing depth and power function or some other criteria such as FDR, and decide sequencing depth according to the "detection power" you want in term of the criterion you choose. For example, you can simulate your data via poisson model (reads count of a gene follows poisson distribution) under different sample size and use some tools such as DEGseq to identify differentially expressed genes. Then, you can calculate sensitivity and specificity under different sample sizes and choose a sample size according to the sensitivity and specificity you want.

[mgogol]

What genome?

I don't actually know the answer, but it will probably depend on the size of the genome.