GPrank: an R package for detecting dynamic elements from genome-wide time series

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• High-throughput sequencing (HTS) enables genome-wide time series analysis of various genomic elements.
• Challenges arise from limited time points and replicates in HTS time series data due to high experimental costs.
• Applications include monitoring gene expression (RNA-seq), DNA methylation (BS-seq), and variant frequencies (Pool-seq).

Purpose of the Study:

• To introduce the GPrank R package for robust analysis of genome-wide time series data.
• To address challenges posed by limited sampling and replicates in HTS time series experiments.
• To provide a method for identifying temporally dynamic genomic elements.

Main Methods:

• GPrank utilizes variance information from HTS data pre-processing (probabilistic quantification or beta-binomial model).
• It models time series using two Gaussian process (GP) models: time-dependent and time-independent.
• Bayes factors (BFs) are computed to compare model evidence, ranking genomic elements by temporal dynamics.

Main Results:

• GPrank effectively models short and irregularly sampled time series.
• The package incorporates variance information to prevent false positives without sacrificing computational efficiency.
• Genomic elements are ranked by BFs, enabling identification of the most temporally dynamic regions.

Conclusions:

• GPrank facilitates the detection and visualization of temporally active genomic elements.
• The identified dynamic elements serve as a starting point for downstream analyses.
• This approach enhances understanding of biological processes through improved analysis of HTS time series data.