PLEK

PLEK predicts long non-coding RNAs (lncRNAs) and messenger RNAs (mRNAs) from transcript sequences using an alignment-free k-mer based method to distinguish coding and non-coding transcripts.

Key Features:

• Alignment-Free Approach: Operates without alignment to reference genomes, enabling analysis of transcriptomes from species lacking comprehensive reference genomes.
• Improved k-mer Scheme and SVM: Utilizes an improved k-mer scheme combined with a support vector machine (SVM) algorithm to classify lncRNAs versus mRNAs, with robustness to high indel sequencing errors.
• Evaluation Accuracy: Demonstrated up to 95.6% accuracy on human RefSeq mRNAs and GENCODE lncRNAs using 10-fold cross-validation and achieved over 90% accuracy on other vertebrate datasets.
• Performance Efficiency: Reported runtime approximately eight times faster than CNCI and 244 times faster than CPC on a single thread.
• Sequencing Platform Versatility: Shown to perform well on data from PacBio and 454 sequencing platforms, which are characterized by high indel error rates.

Scientific Applications:

• lncRNA versus mRNA classification: Distinguishes long non-coding RNAs from messenger RNAs within RNA-seq transcriptome datasets.
• Annotation-independent transcript identification: Enables identification of novel transcripts without relying on pre-existing genome annotations.
• Cross-species transcriptome analysis: Facilitates transcriptome studies in species with limited or incomplete genomic resources.

Methodology:

Alignment-free classification using an improved k-mer scheme fed into a support vector machine (SVM), with model evaluation by 10-fold cross-validation on datasets including human RefSeq mRNAs and GENCODE lncRNAs.