Small RNA Transcriptome Profiles of Four Cotton Species, Gossypium hirsutum, G. herbaceum, G. arboreum and G. raimondii

Comparative expression profile analysis of small RNA (sRNA) can aid in the understanding complex regulatory mechanisms. We analyzed the sRNA transcriptomes from Gossypium seeds, leaves and seedlings using bioinformatic tools such as University of East Anglia sRNA (UEA sRNA) tool kit, CLCBio Genomics Workbench, deep-sequencing sRNA analysis pipeline (DSAP) and miRNAKey. We predicted siRNA loci using SiLoCo, and identified siRNA classes using deepBase and Ta-siRNA predictor. Known RNA classes were removed by comparing to Rfam, Noncode and RepBase databases using BLASTN. We identified over 200 conserved miRNA families and novel miRNAs each. Also, several miRNA* sequences, phased ta-siRNAs and different types of siRNAs were identified. The average read lengths of conserved, novel miRNAs and ta-siRNAs identified were 21 nt, 19 nt and 21 nt, respectively as reported earlier. The 5’ terminal of novel mature miRNAs had higher preference to U’s and A’s, a signature of canonical miRNAs suggesting the processing role of AGO1 proteins. Our set of predicted novel cotton miRNA precursors had GC content above 30%, minimum free energy (MFE) lower than -20.0 Kcal/mol, and length ranged from 75 - 310 nt. The predicted gene targets of the novel miRNAs included a diverse set of transcription factors (TFs). The majority of genes or TFs (~90%) we identified were regulated by mRNA cleavage than by translational repression, supporting post-translational gene silencing as predominant microRNA regulatory mechanism. The identification of sRNA as shown here will aid in understanding common regulatory networks in upland cotton G. hirsutum and its progenitors.