Calculating RPM for miRNA

Question

[glady]

Hello, While calculating RPM the read counts should be divided by : total number of MAPPED READS from a given library or by the total number of reads in the sample ?

Thank you.

[ATpoint]

It should be the total number of reads assigned to your list of miRNAs. In a count matrix where columns are samples and rows are genes (or here miRNAs), the total read number per sample would be the column sum.

[Santosh Anand]

Mapped reads, because that is what ultimately defines the depth at which a gene or any genomic element has been sequenced. BTW, use TPM instead of RPKM / FPKM, as the latter does not allow to compare between the samples: C: Which normalization method to use FPKM/TPM?