Long non-coding RNAs in banana: prediction, mapping and their comparative studies using Musa balbisiana and Musa acuminata transcriptome

Abstract

Computational prediction of banana lncRNAs from pooled transcriptome data using 12 different tissues of Musa balbisiana (BB) and Musa acuminata (AA) separately was done and tissue specificity bias was normalized. Musa balbisiana and Musa acuminata are the progenitors of the present day cultivated banana. In this work, we have computationally mined lncRNAs from transcriptome data of M. balbisiana and M. acuminata separately, later we have classified the predicted lncRNAs and identified lncRNAs as Natural antisense RNA or as small ncRNA decoys. The lncRNAs were read mapped to the banana reference genome. 40 randomly selected lncRNAs were validated on three different tissues of M. balbisiana and M. acuminata separately. Here, we report 1319 M. balbisiana lncRNAs and 4288 M. acuminata lncRNAs. We classified the predicted lncRNAs based on its length and we observed that large lncRNAs were present only in M. balbisiana. Among the annotated Musa NATs, 202 were common to the two Musa species and 1397 and 2024 were uniquely present in M. balbisiana and M. acuminata respectively. The read mapping of the lncRNAs revealed that chromosome 8 and 3 harbors maximum lncRNAs in M. balbisiana and M. acuminata respectively. The tissue-specific, as well as species-specific expression of lncRNA was observed through validation of lncRNA by PCR amplification. This study represents a first attempt to mine and compare the differences in lncRNAs between the two Musa species.