FEELnc

FEELnc (FlExible Extraction of LncRNAs) is an alignment-free software program designed to accurately annotate long non-coding RNAs (lncRNAs) from RNA sequencing (RNA-seq) data. RNA-seq has become a foundational technology for exploring RNA populations within biological samples. However, distinguishing between mRNAs (which will be translated) and lncRNAs remains a significant challenge due to the complexity and diversity of the transcriptome.

Key Features of FEELnc include:
- Alignment-Free Annotation: FEELnc does not require alignment to a reference genome, making it versatile and efficient for lncRNA annotation across different species.
- Random Forest Model: The core of FEELnc's classification strategy is a Random Forest model that utilizes general features such as multi k-mer frequencies and characteristics of open reading frames. This approach allows for the accurate differentiation between coding and non-coding transcripts.
- Benchmarking Excellence: When compared against five state-of-the-art tools for lncRNA annotation, FEELnc demonstrates similar or superior performance on benchmark datasets such as GENCODE and NONCODE, highlighting its effectiveness.
- User-Friendly Modules: FEELnc includes specific modules that allow users to adjust classification accuracy, formalize lncRNA class annotations, and identify lncRNAs even without a pre-existing set of non-coding RNAs for training, enhancing its flexibility and utility.
Comprehensive Annotation: Applying FEELnc to a dataset of 20 canine RNA-seq samples from the European LUPA consortium significantly expanded the canine genome annotation, adding 10,374 novel lncRNAs and 58,640 mRNA transcripts, thereby demonstrating its capacity to enhance genomic resources.