Positional correlation analysis improves reconstruction of full-length transcripts and alternative isoforms from...

Area of Science:

• Transcriptomics
• Bioinformatics
• Genomics

Background:

• Reconstructing full-length transcripts is crucial for understanding transcriptome complexity.
• Existing methods struggle with high noise and biases from sequencing data.
• A robust method is needed for accurate transcript reconstruction across different platforms.

Purpose of the Study:

• To develop a novel statistical method for accurate full-length transcript reconstruction.
• To create a method robust against noise and biases inherent in transcriptome data.
• To enable transcript reconstruction from both next-generation sequencers and tiling arrays.

Main Methods:

• ARTADE2 analyzes 'positional correlation' across genomic positions in multiple transcriptional datasets.
• A logistic model combined with a Markov model is used for transcript reconstruction.
• The method incorporates factor analysis to detect potential alternative splicing regions.

Main Results:

• ARTADE2 reconstructed 17,591 full-length transcripts from 55 datasets, outperforming existing methods.
• 1489 novel transcripts were discovered, with 14 experimentally validated.
• The method demonstrated strong performance in reconstructing mRNA from next-generation sequencing data.

Conclusions:

• ARTADE2 offers a robust and accurate approach to full-length transcript reconstruction.
• The method's ability to identify novel transcripts and potential isoforms has significant implications for biomarker discovery.
• ARTADE2 is applicable across various sequencing technologies, enhancing its utility in diverse research fields.