Difficulty in getting Novel miRNAs

[vivekruhela]

I am trying to identify miRNA from my data. In my data, I knew apriorly that sample contains novel miRNAs and I am also getting some of them when using mirdeep2/mirdeep* sequence aligner. But I am not getting any novel miRNa with miRge2.0 pipeline. I don't understand why this is happening. I am attaching here some of the sequences which are novel miRNAs but not determined by miRge2.0. Can you suggest me why this is happening or did I missing anything?

Novel miRNA_sequence

22 uuggaguaguauuccucuaggccc
23 uuccccagccuugucuagaca
24 uggcuggcugcuccgggcac
25 gugugugcaccugugucugucug
26 caggcgcggacuccacacccagc
27 caagagccuggaacugcaccaa
28 caagagccuggaacugcaccaa
29 guccccagacucucccgggcug
30 acagggcagccuggcaccugc
31 caagagccuggaacugcaccaa
32 caagagccuggaacugcaccaa
33 acaggacaagggcagccccacc
34 caagagccuggaacugcaccaa
35 ccggcugcggcucccaccu
36 aaaaguaauuguuguucuugcc
37 gucccuguucgggcgcca
38 uacccugacugucccucuguaga

Thanks.

[mhalushka]

Hi. I don't have a specific answer without knowing your exact data. In general, the miRge2.0 pipeline, as described in the paper, is much more specific than miRDeep2 in calling novel miRNAs. miRDeep2 novel miRNAs are mostly not bonafide miRNAs, just short sequences. Bastian Fromm has several papers on the topic. miRge2.0 novel miRNA caller was dependent on a number of variables including the presence of 5p and 3p sequences, abundance of reads, the structure of the reads to each other, etc. My suspicion is that those listed novel miRNAs above didn't get picked up because of some aspect of what I just described was missing. We designed our tool to really only identify high-confidence miRNAs. If you think you have real miRNAs that are being missed by miRge, but found by miRDeep2, then use that information.