microbiomeDASim

microbiomeDASim simulates longitudinal differential abundance in microbiome datasets as an R Bioconductor package to support evaluation of statistical methods for longitudinal metagenomic studies.

Key Features:

• Longitudinal Data Simulation: Generates synthetic longitudinal microbiome feature data that mimic temporal dynamics observed in longitudinal metagenomic sampling.
• Flexible Functional Forms: Allows specification of various functional forms for mean trends with flexible parameters to model different biological scenarios.
• Multivariate Normal Model: Draws observations from a multivariate normal distribution to represent realistic variability and correlation structures.
• Signal-to-Noise Control: Provides parameters to control the signal-to-noise ratio for precise manipulation of detectable effects in simulated data.

Scientific Applications:

• Methodological Comparison: Simulates data under varied conditions to compare performance of statistical methods for detecting differential abundance over time.
• Study Design and Planning: Models potential outcomes to aid study design and planning for longitudinal metagenomic experiments, including considerations of cost and logistical constraints.
• Framework Evaluation: Evaluates different statistical frameworks in scenario-specific simulations to inform method selection for particular datasets.

Methodology:

Observations are generated from a multivariate normal distribution using specified functional forms with flexible parameters to create controlled longitudinal datasets; the package highlights metaSplines as an example method for differential abundance detection.