SimSeq

SimSeq simulates RNA-seq read count datasets nonparametrically to generate data whose joint read-count distribution matches user-provided experimental RNA-seq datasets for evaluation of RNA-seq analysis methods.

Key Features:

• Nonparametric simulation algorithm: Uses a data-driven, nonparametric approach rather than parametric models such as the negative binomial to simulate read counts.
• Realistic data generation: Produces simulated datasets that capture the variability and complexity of user-provided experimental RNA-seq data by matching observed distributions.
• Performance benchmarking and FDR assessment: Enables comparison of RNA-seq analysis methods, including evaluation of false discovery rate (FDR) control under parametric and nonparametric simulation scenarios.

Scientific Applications:

• Method validation: Validates differential expression and other RNA-seq analysis methods using simulated data that reflect real experimental distributions.
• Benchmarking: Benchmarks algorithm performance and FDR control under realistic, data-driven simulation scenarios versus parametric simulations.
• Statistical method development: Supports development and assessment of statistical methods robust to the complexities of real-world RNA-seq read counts.

Methodology:

SimSeq applies a nonparametric, data-driven simulation algorithm that leverages user-provided RNA-seq datasets to match the joint distribution of read counts and generate simulated data for evaluating analysis methods and FDR control.