Hi Swati

You can sequence the samples separately using barcoded libraries in a single lane. While analyzing you can treat the files together as a single file for identifying novel miRNA or separately as biological replicates to do some statistical analysis while calculating differential expression of the miRNAs. I prefer to sequence the biological replicates separately.

If you pool samples, then your single sample gives an approximation of the mean abundance of each miRNA. If you sequence each replicate separately, then you can estimate the mean by combining the results, but you can also estimate the biological variance between samples.

The biological variation between replicates is necsessary information if you want to compare between treatments, so there is a definite advantage to NOT pooling.

The advantage of pooling is that you will be able to sequence the sample to greater depth for the same cost. This might mean an increased chance to observe rare transcripts, which would have more chance to be missed in the shallower sequencing of the three individual samples.

Indeed, there is two aspects of your question that needs to be clarified:

- you can "pool" different samples on a single lane, by barcoding them differently. In that case, the only consideration is the depth of coverage of each sample: one sample per lane will give you more reads than 3 samples per lane. Of course the higher the better to identify rare transcripts, but then it is a matter of cost.

-pooling or not pooling your biological replicates, prior to library construction and RNA-Seq. In that case the issue is biological variability and as Jonathan said you should NOT pool bio reps. Otherwise, no possible comparisons, no stats and ultimately, no publication (at least in good journals).

Completely Agree with Jean.