FEELnc: a tool for long non-coding RNA annotation and its application to the dog transcriptome

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• Whole transcriptome sequencing (RNA-seq) is crucial for RNA analysis.
• Distinguishing long non-coding RNAs (lncRNAs) from messenger RNAs (mRNAs) remains a challenge.
• Accurate annotation of RNA classes is essential for understanding gene regulation.

Purpose of the Study:

• To introduce FEELnc, a novel alignment-free program for accurate lncRNA annotation.
• To provide a flexible and standardized solution for classifying RNA types.
• To expand genome annotation using RNA-seq data.

Main Methods:

• Developed FEELnc, an alignment-free program utilizing a Random Forest model.
• Incorporated features like k-mer frequencies and relaxed open reading frames for classification.
• Benchmarked FEELnc against five state-of-the-art tools using GENCODE and NONCODE datasets.
• Included modules for fine-tuning accuracy and annotating lncRNA classes without a training set.

Main Results:

• FEELnc demonstrated comparable or superior performance to existing tools.
• Applied FEELnc to canine RNA-seq data, identifying 10,374 novel lncRNAs.
• Discovered 58,640 novel mRNA transcripts in the canine genome.
• Successfully expanded canine genome annotation significantly.

Conclusions:

• FEELnc offers a standardized and comprehensive approach to lncRNA annotation.
• The tool effectively distinguishes lncRNAs from mRNAs, improving RNA-seq data interpretation.
• FEELnc facilitates substantial expansion of genome annotations, particularly for non-model organisms.