mmannot

mmannot quantifies small non-coding RNAs such as microRNAs (miRNAs), tRNA-derived fragments, small nucleolar RNAs (snoRNAs), and small nuclear RNAs (snRNAs) from high-throughput RNA-seq by evaluating multi-mapping reads against genomic annotations to produce unbiased feature-level counts.

Key Features:

• Feature quantification: Targets the feature quantification step of high-throughput RNA-seq studies to assign reads to annotated features.
• Supported RNA classes: Processes small non-coding RNA classes including miRNAs, tRNA-derived fragments, snoRNAs, and snRNAs.
• Reference alignment comparison: Uses genomic read alignments and compares read positions with existing annotations to assign features.
• Multi-mapping evaluation: Evaluates all reads that map to multiple genomic locations and inspects their associated annotations rather than discarding or arbitrarily assigning them.
• Co-localization handling: Detects hits that co-localize with identical feature annotations, for example duplicated miRNAs or gene copies, and accounts for them in counts.
• Annotation merging: Merges different annotations when a read maps to distinct features and reports counts for the newly formed merged entities.
• Bias reduction: Provides an alternative to conventional multi-mapping resolution strategies to reduce bias in feature quantification.

Scientific Applications:

• sRNA quantification from RNA-seq: Quantifies genome-wide distributions of miRNAs, tRNA-derived fragments, snoRNAs, and snRNAs from RNA-seq experiments.
• Ambiguity resolution: Improves accuracy of feature-level counts by resolving ambiguities from multi-mapping reads.
• Duplicated feature analysis: Enables detection and combined quantification of duplicated miRNAs or gene copies through co-localization and merged-feature counts.
• Cross-organism sRNA profiling: Supports profiling of small RNA distributions across different organisms using genome-wide sequencing data.

Methodology:

Reads are aligned to a reference genome and compared with existing annotations; all multi-mapping reads and their associated annotations are evaluated, co-localized identical annotations are recognized, and when reads map to different annotations those annotations are merged and counted as combined features.