A comprehensive annotation and differential expression analysis of short and long non-coding RNAs in 16 bat genomes

Nelly Mostajo Berrospi; Marie Lataretu; Sebastian Krautwurst; Florian Mock; Daniel Desirò; Kevin Lamkiewicz; Maximilian Collatz; Andreas Schoen; Friedemann Weber; Manja Marz; Martin Hölzer

If not otherwise stated, the main ncRNA-discovery is based on homology searches against the Rfam database 46-48 (v14.0). We used the Gorap pipeline (https://github.com/koriege/gorap), a specially developed software suite for genomewide ncRNA screening. Gorap screens genomic sequences for all ncRNAs present in the Rfam database using a generalized strategy by applying multiple filters and specialized software tools. To this end, Gorap takes huge advantage of Infernal (v1.1.2) 49,50 to annotate ncRNAs based on input alignment files conserved in sequence and secondary structure (so-called Stockholm alignment files; stk). All resulting alignment files were automatically pre-filtered by Gorap based on different ncRNA class-specific parameters including taxonomy, secondary structure, and primary sequence comparisons. Due to repeats, pseudogenes, undiscriminable un-/functional genes, and overlapping results from the different assembly methods, we defined a ncRNA set per species for annotation that includes filtered sequences, but allows for variants and multiple copies. This final annotation set is defined by hand-curating the resulting stk alignments of Gorap with the help of Emacs RALEE mode 51 . Due to the removal of sequences in the stockholm alignments, the remaining sequences were extracted and again aligned into stockholm format using cmalign -noprob from Infernal. The Rfamderived ncRNA alignments were further split into snoRNAs (Tab. S5), miRNAs (Tab. S6), and other ncRNAs including snRNAs, lncRNAs, and other structural RNAs (Tab. S8).