RAmpSim: a thermodynamic simulator for hybridization capture in metagenomic sequencing

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Synthetic datasets are crucial for developing and benchmarking computational tools due to the lack of ground truth.
• Existing read simulators often assume uniform genome sampling, which is not suitable for capture-based sequencing technologies like TELSeq.
• Capture-based sequencing introduces systematic biases (probe multiplicity, sequence composition, species abundance) and intentionally oversamples regions of interest, necessitating a specialized simulator.

Purpose of the Study:

• To develop a novel simulator, RAmpSim, specifically designed for capture-based sequencing data.
• To accurately model the unique characteristics and biases of capture-based sequencing technologies.
• To provide a tool that improves the accuracy of synthetic data generation for benchmarking computational methods.

Main Methods:

• RAmpSim models bait-target hybridization and fragment capture using a thermodynamic nearest-neighbor energy model.
• It employs Boltzmann-weighted sampling of binding sites and multinomial sampling for fragment generation, parameterized by bait concentration, binding energy, and genomic abundance.
• The simulator integrates platform-specific error models and is implemented in the Rust programming language.

Main Results:

• RAmpSim accurately reproduces empirical within-genome coverage and cross-species enrichment patterns observed in capture-based metagenomic datasets.
• It generally outperforms a uniform baseline in terms of position-based earth mover's distance when compared against empirical coverage distributions.
• Classification analysis demonstrates RAmpSim's high recall in recovering empirical high-coverage regions, surpassing uniform baselines.

Conclusions:

• RAmpSim is an effective and fast simulator for capture-based sequencing data.
• The simulator accurately captures the complexities and biases inherent in capture-based sequencing, outperforming uniform sampling models.
• RAmpSim provides a valuable resource for developing and validating computational tools used in genomics and metagenomics.