DE analysis using mature miRNAs with different precursor sequences

[cmorgan]

Hi,

I have an interesting situation. I am trying to perform a DE analysis on miRNAseq data using mirDeep2 from the miRNAs_expressed_all_samples_now.csv file.

I am interested in the DE of my mature miRNA sequences however, there are sevreal cases where my mature miRNA has a different precursor sequence, for example:

mmu-let-7a-5p 54439.00 mmu-let-7a-1 54439.00 54439.00 5414.39
mmu-let-7a-5p 54422.00 mmu-let-7a-2 54422.00 54422.00 5412.70

The expression values are similar but not the same. Has anyone dealt with this before and could you give me some advice? I am confused on whether i should (i) treat these mature miRNAs with different precursors as unique expression values (ii), if should use expression values from just one of these miRNAs or (iii), if i should add the expression values from both entires of mmu-let-7a-5p to get the actual expression value.

Thanks in advance for any assistance with this query.

[gpcr]

mmu-let-7a-1
chr13: 48538179-48538272
5'-UUCACUGUGGGAUGAGGUAGUAGGUUGUAUAGUUUUAGGGUCACACCCACCACUGGGAGAUAACUAUACAAUCUACUGUCUUUCCUAAGGUGAU-3' 94

mmu-let-7a-2
chr9: 41536716-41536811
5'-CUGCAUGUUCCCAGGUUGAGGUAGUAGGUUGUAUAGUUUAGAGUUACAUCAAGGGAGAUAACUGUACAGCCUCCUAGCUUUCCUUGGGACUUGCAC-3' 96


mmu-let-7a-5p has same sequence coming from two different precursors located at different chromosome as seen above.
Since mature sequence is same, origin of it is difficult to know; mirdeep2 tries to fit the mature sequence into both the precursors.

My experience says that we should use mmu-let-7a-5p as only once. I follow to take only one with highest numbers for DE analysis. otherwise same count would be represented twice which skews the data.