Fundamental question about biological replicates and cuffdiff

Question

[halo22]

I will soon start working with RNA-seq for analyzing gene expression for patient groups. I understand that having both technical and biological replicates are important in designing a RNA-seq experiment. Since I will mostly be analyzing samples from healthy and diseased individuals, I don't think I will have any technical replicates as the samples are from different individuals and sequenced separately. Is it safe to say that I will have 9 biological replicates per condition, assuming that I have 10 individuals from each group?

Also, I was going through the cuffdiff documentation, could somebody please explain how the p-value is calculated for the gene-expression file?

Thanks

Answer 1

[Devon Ryan]

Technical replicates have little if any value in RNAseq. Your different patient samples are biological replicates if they belong to the same group (e.g., "treated" or "diseased"). If you have 10 samples in a group, then you have 10 biological replicates.

Regarding cuffdiff, you're be hard pressed to find anyone that actually knows how the current version works internally, it's more or less a black box (unless someone really wants to go through the source code). In general, you're better off with DESeq2, edgeR, or limma (they're the best practice).