cleanUpdTSeq

cleanUpdTSeq applies a Naive Bayes classifier (e1071 R package) trained on zebrafish data to classify and filter putative polyadenylation (pA) sites in high-throughput sequencing data, reducing oligo(dT) internal-priming artifacts to improve poly(A) site identification.

Key Features:

• Naive Bayes classifier (e1071): Utilizes a Naive Bayes classifier implemented via the e1071 R package to assign probability scores to candidate pA sites.
• Zebrafish-derived training data: Employs zebrafish training datasets to parameterize and inform the classifier.
• Internal priming artifact reduction: Targets and reduces false positives caused by oligo(dT) internal priming in 3' end sequencing data.
• 3'READS compatibility: Designed to operate on data generated by modified 3'READS (3' region extraction and deep sequencing) protocols to leverage high-quality 3' end reads.
• Probabilistic scoring for isoform analysis: Provides probabilistic evaluations of pA sites to support downstream quantification of 3' mRNA isoforms.

Scientific Applications:

• Gene expression regulation studies: Enables accurate identification and quantification of 3' mRNA isoforms to study alternative polyadenylation in gene regulation.
• Genome-wide poly(A) site mapping: Facilitates comprehensive mapping of poly(A) sites across genomes for transcriptomics and post-transcriptional regulation research.
• Quantitative analysis of mRNA isoforms: Supports measurement of relative abundances of 3' mRNA isoforms to inform functional and regulatory studies.

Methodology:

Processes reads compatible with modified 3'READS protocols and applies a Naive Bayes classifier implemented in the e1071 R package, trained on zebrafish-derived data, to probabilistically score and filter candidate polyadenylation sites and minimize oligo(dT) internal-priming artifacts.