Classification of real and pseudo microRNA precursors using local structure-sequence features and support vector machine

Abstract

BackgroundMicroRNAs (miRNAs) are a group of short (~22 nt) non-coding RNAs that play important regulatory roles. MiRNA precursors (pre-miRNAs) are characterized by their hairpin structures. However, a large amount of similar hairpins can be folded in many genomes. Almost all current methods for computational prediction of miRNAs use comparative genomic approaches to identify putative pre-miRNAs from candidate hairpins. Ab initio method for distinguishing pre-miRNAs from sequence segments with pre-miRNA-like hairpin structures is lacking. Being able to classify real vs. pseudo pre-miRNAs is important both for understanding of the nature of miRNAs and for developing ab initio prediction methods that can discovery new miRNAs without known homology.ResultsA set of novel features of local contiguous structure-sequence information is proposed for distinguishing the hairpins of real pre-miRNAs and pseudo pre-miRNAs. Support vector machine (SVM) is applied on these features to classify real vs. pseudo pre-miRNAs, achieving about 90% accuracy on human data. Remarkably, the SVM classifier built on human data can correctly identify up to 90% of the pre-miRNAs from other species, including plants and virus, without utilizing any comparative genomics information.ConclusionThe local structure-sequence features reflect discriminative and conserved characteristics of miRNAs, and the successful ab initio classification of real and pseudo pre-miRNAs opens a new approach for discovering new miRNAs.

Application of massively parallel sequencing to microRNA profiling and discovery in human embryonic stem cells

“…Small RNA library preparation Undifferentiated H9 hESCs (Hirst et al 2007) were cultured on Matrigel (BD Biosciences) coated dishes in maintenance medium consisting of Dulbecco's Modified Eagle Medium (DMEM)/F12 containing 20% Knockout Serum Replacer (Invitrogen), 0.1 mM ␤-mercaptoethanol, 0.1 mM nonessential amino acids, 1 mM glutamine, and 4 ng/mL FGF2 (R&D Systems) and conditioned by mitotically inactivated mouse embryonic fibroblasts (Xue et al 2005). For EB differentiation, hESCs were harvested via 0.05% trypsin (Invitrogen) supplemented with 0.5 mM CaCl 2 , and the resultant cell aggregates were cultured in nonadherent dishes (BD Biosciences) for up to 30 d in maintenance medium lacking FGF2 (medium changes performed as necessary) (Itskovitz-Eldor et al 2000;Dvash et al 2004).…”

Computational identification of novel microRNAs and their targets in the malarial vector, Anopheles stephensi

“…Validation of the identified secondary structures was performed using M-Fold tool based on thermodynamics metrics (Zuker 2003). The annotation of mature miRNA was performed based on the widely accepted criteria for selecting miRNAs from pre-miRNA sequences (Ambros et al 2003;Xue et al 2005). These were (a) presence of mature miRNA sequence of at least 18 nucleotide length on the stem including GU Wobble pairs (b) Minimum Free energy value \ -15 kcal/mol (c) Absence of multiple loops (d) ability to fold to stem-loop/hair-pin structures.…”

Pse-in-One 2.0: An Improved Package of Web Servers for Generating Various Modes of Pseudo Components of DNA, RNA, and Protein Sequences

“…1) The occurrences of kmers, allowing at most m mismatches (Mismatch) [264,265,292] 2) The occurrences of kmers, allowing non-contiguous matches (Subsequence) [265,292,293] Autocorrelation 3) Moran autocorrelation (MAC) [217,294] 4) Geary autocorrelation (GAC) [217,295] 5) Normalized Moreau-Broto autocorrelation (NMBAC) [217,296] Predicted structure composition 6) Local structure-sequence triplet element (Triplet) [266] 7) Pseudo-structure status composition (PseSSC) [226] 8) Pseudo-distance structure status pair composition (PseDPC) [10] 2) PseAAC of Distance-Pairs and Reduced Alphabet (Distance Pair) [271] Autocorrelation 3) Physicochemical distance transformation (PDT) [270] Profile-based features 4) Select and combine the n most frequenct amino acids according to their frequencies (Top-n-gram) [269] 5) Profile-based Physicochemical distance transformation (PDT-Pofile) [270] 6) Distance-based Top-n-gram (DT) [271] 7) Profile-based Auto covariance (AC-PSSM) [272] 8) Profile-based Cross covariance (CC-PSSM) [272] 9) Profile-based Auto-cross covariance (ACC-PSSM) [272] Natural Science Mismatch [264] and Subsequence [265]; and 3 are added into the autocorrelation category, i.e., Moran autocorrelation, Geary autocorrelation, and Normalized Moreau-Broto autocorrelation [268]. PseAAC-General is designed to generate the feature vectors for protein sequences.…”