ppcseq

ppcseq identifies transcripts containing outlier data points in differential expression analysis that deviate from a negative binomial distribution.

Key Features:

• Probabilistic outlier detection: Uses Bayesian computation to compare observed RNA sequencing data against its theoretical negative binomial distribution for large-scale outlier identification.
• Quality control for differential expression: Flags transcripts with extreme outliers that inflate differential expression statistics in RNA sequencing generalized linear models.
• Cross-algorithm compatibility: Can be applied to outputs from multiple popular differential expression analysis tools and public datasets.

Scientific Applications:

• RNA sequencing differential expression quality control: Identifies 3 to 10 percent of differentially abundant transcripts whose statistics are inflated by outliers across algorithms and datasets.
• Negative binomial model validation: Detects transcripts that violate negative binomial assumptions in transcript abundance analysis.

Methodology:

Bayesian computation is used for large-scale comparison of observed data against its theoretical distribution given in a statistical model.