miRNA identification, mapping and quantification
Detection of known and novel miRNAs was done using miRDeep2 (Friedländer et al., 2008; Friedländer, Mackowiak, Li, Chen, & Rajewsky, 2012). First, reads from all samples were merged and mapped to a P. napigenome (Lohse, Hayward, Ebdon, of Life, & Consortium, 2021). Reads were compared to known miRNAs from all mature miRNAs and hairpin sequences from B. mori and all mature reads from Heliconius melpomene from miRbase v22 (Kozomara, Birgaoanu, & Griffiths-Jones, 2019). Remaining unmapped sequences were then used to identify novel miRNAs and their hypothetical secondary structure (Friedländer et al., 2012), giving an output of known and novel miRNAs in all samples in the data set. Identified miRNAs were then discarded if they did not meet threshold requirements of a miRDeep2 score ≥ 1, presence in more than 5 samples, and a total read coverage of ≥ 10 reads across samples. All identified miRNAs were then used to obtain read counts for each miRNA for each sample using the quantifier.pl script in miRDeep2.