UMI-Gen

UMI-Gen simulates paired-end Next Generation Sequencing (NGS) reads with Unique Molecular Identifiers (UMIs) to generate realistic datasets for benchmarking and evaluating low-frequency single nucleotide variant (SNV) and copy number variation (CNV) detection in cancer genomics.

Key Features:

• UMI Integration: Incorporates Unique Molecular Identifiers (UMIs) to distinguish true genetic variants from sequencing and PCR artifacts.
• Paired-end Read Simulation: Generates reference paired-end reads covering targeted regions.
• Customizable Depths: Produces simulated reads at user-defined coverage depths to model various experimental conditions.
• Background Error Rate Estimation: Uses control files to estimate background error rates at each reference position and incorporate those errors into simulated reads.
• Real Variant Insertion: Inserts user-provided lists of known variants into the simulated dataset.
• Data Characteristic Modeling: Modifies generated reads to reflect real biological and sequencing data characteristics.
• Benchmarking Capability: Produces datasets that resemble real biological samples to benchmark UMI-based variant callers, including evaluation of variants at frequencies below 1%.

Scientific Applications:

• Benchmarking UMI-based Variant Callers: Evaluates sensitivity and specificity of variant callers using controlled simulated datasets.
• Low-frequency Variant Evaluation in Cancer Genomics: Assesses detection performance for low-frequency SNVs and CNVs relevant to cancer studies.

Methodology:

Generates reference paired-end reads for targeted regions at user-defined depths; estimates background error rates per position using control files; modifies reads to reflect sequencing and PCR error characteristics; inserts user-provided variants into the simulated dataset.