Bygul: Amplicon & Metagenomics Read Simulator

Bygul is a Python 3 tool designed for simulating sequencing reads in wastewater surveillance and other metagenomic applications. It allows users to simulate complex multi-sample datasets with customizable proportions using industry-standard backends like wgsim and mason.

---

🏗 Installation

Bygul requires Python 3. Since it relies on external simulators (wgsim and mason), we recommend using Conda to manage dependencies.For more info on wgsim and mason simulator please check their documentations.

Option 1: Via Conda (Recommended)

conda create -n bygul bioconda::bygul

Option 2: Via PyPI

pip install bygul

Note: Some binary dependencies (wgsim/mason) may need to be installed manually or built from source if using this method.

Option 3: Local Build from Source

git clone [https://github.com/andersen-lab/Bygul](https://github.com/andersen-lab/Bygul)
cd Bygul
pip install -e .

---

🧬 Usage: Amplicon Sequencing Mode

Use this mode when simulating specific genomic regions defined by a primer set.

Basic Command

bygul simulate-proportions --genomes [SAMPLE1.fasta,SAMPLE2.fasta] --primers [primer.bed] --reference [reference.fasta] --proportions [0.8,0.2] --outdir [output_dir]

Advanced Examples

• Random Proportions & Mismatches: Simulate with random proportions and allow up to 2 SNPs in primer regions.

  bygul simulate-proportions --genomes sample1.fasta,sample2.fasta --primers primer.bed --reference reference.fasta --outdir results/ --maxmismatch 2

• Switching Simulators: Use mason instead of the default wgsim.

  bygul simulate-proportions --genomes sample1.fasta,sample2.fasta --primers primer.bed --simulator mason

• Custom Error Rates & Lengths: Pass simulator-specific parameters (e.g. indel fraction -R) directly.

  bygul simulate-proportions --genomes sample1.fasta,sample2.fasta --primers primer.bed -R 0.01

• Using a csv file and all samples in a multi-fasta file:

  bygul simulate-proportions --csv samples.csv --multifasta samples.fasta

---

🌍 Usage: Metagenomics Mode

Simulate reads from entire samples without requiring a primer BED file or a reference sequence.

Basic Metagenomics Simulation

bygul simulate-proportions sample1.fasta,sample2.fasta --outdir results/ --simulation_mode metagenomics

Metagenomics with Specific Parameters

bygul simulate-proportions sample1.fasta,sample2.fasta --proportions 0.5,0.5 --outdir results/ --simulation_mode metagenomics --simulator mason --illumina-read-length 200

Metagenomics with csv and multifasta

bygul simulate-proportions --csv samples.csv --multifasta samples.fasta --outdir results/ --simulation_mode metagenomics

---

📝 Technical Notes

Parameter Handling

Bygul acts as a wrapper. While most flags are passed directly to the underlying simulators, the following are managed directly by Bygul for more realistic simulations(amplicon simulation mode only):

• --readcnt: Number of reads per amplicon.
• --wgsim_insert_size: Insert size for wgsim.
• --wgsim_read_length / --wgsim_error_rate.

To see all available backend flags, run:

wgsim --help
mason_simulator --help

Please note that some dependencies are not available through pypi. You need to install them using conda or build from source.

Reference file

Reference file is used only when the provided bed file does not have the sequence column. We strongly recommend for your file to have a sequence column as the program will extract the sequences from the reference sequence if not provided.

Using a CSV file for sample names and proportions

--csv and --multifasta are always provided together, the CSV file contains two columns sample_name and proportion. Samples with multiple contigs, must have their IDs as: sample_name|contig_name in the multifasta file.

Number of reads per amplicon

It is recommended to define the number of reads per amplicon to be greater than the number of contigs in your amplicon file. This is particularly important when your primers are designed for whole genome sequencing, where each amplicon may contain a substantial number of contigs. Setting too few reads per amplicon may result in empty read files for certain amplicons, leading to incomplete simulated reads.

Primer bed file

🧬 Input BED File Format

The pipeline expects a tab-delimited BED file where the first six columns represent standard genomic coordinates (chrom, chromStart, chromEnd, name, poolName, strand). Crucially, the fourth column (name) must follow a strict naming convention to prevent downstream parsing failures in variant-calling tools: [Scheme-Name]_[AmpliconNumber]_[Direction]_[OptionalSuffix] (e.g., SARS-CoV-2_3_LEFT or SARS-CoV-2_3_LEFT_alt). To ensure structural boundaries are parsed correctly, the prefix must not contain underscores, and any optional trailing modifiers must be restricted to standard alternative tags (_alt, _ALT1) or tracking indexes (_0, _1). Multi-level pool formatting, such as SARS-CoV-2_400_1_LEFT_1, is malformed and will fail validation.. The maximum number of mismatches allowed for each primer sequence is 1 SNP. To change this number, you may use the --maxmismatches flag.

Complete set of available parameters

To learn more about how to adjust other parameters for the simulator please read the documentation for wgsim and mason simulator. Users can pass any simulator parameter directly in their command. The only parameters set through bygul are --readcnt and --wgsim_insert_size,--wgsim_read_length and --wgsim_error_rate.

Simulated reads output

Simulated reads from all samples are located in provided_output_path/reads_1.fastq and provided_output_path/reads_2.fastq

Information about amplicon dropouts

In order to find out more about amplicon dropouts, please refer to provided_output_path/sample_name/amplicon_stats.csv file. Please note that primer_seq_x and primer_seq_y define the left and right primer sequence whereas left_match and right_match shows the actual sequence found in the sample for a better comparison of mismatching bases in the primer sequence. Additionally, if there are any ambiguous bases present in the matching sequence, the ambiguous_bases value returns true.