Many of the library prep kits provide information on the typical number of reads required for analysis. I know this can depend on the type of analysis but I was hoping to get a bit more clarity on terminology. Using Illuminia as an example, they say "Gene expression profiling experiments that are looking for a quick snapshot of highly expressed genes may only need 5 million to 25 million reads per sample. In these cases, researchers can pool multiple RNA-Seq samples into one lane of a sequencing run, which allows for high multiplexing of samples". So if we say we need a minimum of 5 million reads per sample. For RNA libraries that have been multiplexed and sequenced using multiple flowcell lanes, should the combined number of reads from all the technical replicates have at least 5 million reads (if they are merged), or each replicate should have 5 million reads after sequencing?
So if we say we need a minimum of 5 million reads per sample.
While you can sequence any amount, if you want to get a good representation of expression the number would be dependent on genome/transcriptome size. 5 M should be enough for bacterial genomes but not for human/mouse genome.
should the combined number of reads from all the technical replicates
have at least 5 million reads (if they are merged), or each replicate
should have 5 million reads after sequencing?
Technical replicates of sequencing (at least for Illumina) are stable so the total number of reads should be the recommended/desired number per sample (could be on multiple lanes as a part of large pool). Reads for each biological replicate on the other hand will need to be "N" million as noted above.