misha (version 5.4.0)

gsynth.random: Generate random genome sequences

Description

Generates random DNA sequences based on nucleotide probabilities without using a trained Markov model. Each nucleotide is sampled independently according to the specified probabilities.

Usage

gsynth.random(
  intervals = NULL,
  output_path = NULL,
  output_format = c("misha", "fasta", "vector"),
  nuc_probs = c(A = 0.25, C = 0.25, G = 0.25, T = 0.25),
  mask_copy = NULL,
  seed = NULL,
  n_samples = 1,
  iterator = 1
)

Value

When output_format is "misha" or "fasta", returns invisible NULL and writes the random sequences to output_path. When output_format is "vector", returns a character vector of sequences (length = n_intervals * n_samples).

Arguments

intervals

Genomic intervals to sample. If NULL, uses all chromosomes.

output_path

Path to the output file (ignored when output_format = "vector")

output_format

Output format:

• "misha": .seq binary format (default)

• "fasta": FASTA text format

• "vector": Return sequences as a character vector (does not write to file)

nuc_probs

Nucleotide probabilities. Can be specified as:

• A named vector: c(A = 0.3, C = 0.2, G = 0.2, T = 0.3)

• An unnamed vector in A, C, G, T order: c(0.3, 0.2, 0.2, 0.3)

Probabilities are automatically normalized to sum to 1. Default is uniform (0.25 each).

mask_copy

Optional intervals to copy from the original genome instead of random sampling. Use this to preserve specific regions exactly as they appear in the reference.

seed

Random seed for reproducibility. If NULL, uses current random state.

n_samples

Number of samples to generate per interval. Default is 1.

iterator

Iterator for position resolution. Default is 1 (base-pair resolution). Larger values may speed up processing but are typically not needed for random sampling.

Details

Unlike gsynth.sample which uses a trained Markov model to generate sequences that preserve k-mer statistics, gsynth.random generates purely random sequences where each nucleotide is sampled independently. This is useful for generating baseline random sequences or sequences with specific GC content.

Nucleotide ordering: When using an unnamed vector for nuc_probs, the order is A, C, G, T. Named vectors can be in any order.

See Also

gsynth.sample, gsynth.train

Examples

gdb.init_examples()

# Generate random sequences with uniform nucleotide probabilities
seqs <- gsynth.random(
    intervals = gintervals(1, 0, 1000),
    output_format = "vector",
    seed = 42
)

# Generate GC-rich sequences (60% GC)
gc_rich <- gsynth.random(
    intervals = gintervals(1, 0, 1000),
    output_format = "vector",
    nuc_probs = c(A = 0.2, C = 0.3, G = 0.3, T = 0.2),
    seed = 42
)

# Generate AT-rich sequences
at_rich <- gsynth.random(
    intervals = gintervals(1, 0, 1000),
    output_format = "vector",
    nuc_probs = c(A = 0.35, C = 0.15, G = 0.15, T = 0.35),
    seed = 42
)