Struct IlluminaErrorModel 

pub struct IlluminaErrorModel { /* private fields */ }

Illumina-style position-dependent error model.

Error rate starts at min_error_rate for early cycles and linearly ramps to max_error_rate by the last cycle. The ramp begins at approximately 70% of the read length, mimicking the quality decay caused by phasing and dephasing in Illumina sequencing.

Read 2 has a higher error rate than read 1 (controlled by r2_multiplier), reflecting the typically worse quality of the second read in a pair.

Error probabilities and base quality scores are precomputed at construction time for each cycle and read end, so the per-base hot path is a table lookup rather than repeated floating-point math.

Implementations

impl IlluminaErrorModel

pub fn new(read_length: usize, min_error_rate: f64, max_error_rate: f64) -> Self

Create a new Illumina error model with default R2 multiplier (1.5x).

Arguments

• read_length — Number of cycles (bases) per read.
• min_error_rate — Per-base error rate at the start of reads.
• max_error_rate — Per-base error rate at the end of reads.

Panics

Panics if read_length is 0 or if min_error_rate > max_error_rate.

pub fn with_r2_multiplier( read_length: usize, min_error_rate: f64, max_error_rate: f64, r2_multiplier: f64, ) -> Self

Create a new Illumina error model with a custom R2 multiplier.

The r2_multiplier scales the entire error curve for read 2 relative to read 1. A value of 1.0 means R1 and R2 have identical error rates.

Panics

Panics if read_length is 0 or if min_error_rate > max_error_rate.

Trait Implementations

impl ErrorModel for IlluminaErrorModel

fn error_probability(&self, cycle: usize, read_end: ReadEnd) -> f64

Return the probability of a sequencing error at the given cycle position (0-based) for the given read end.

fn base_quality_phred33(&self, cycle: usize, read_end: ReadEnd) -> u8

Return the Phred+33 base quality score for the given cycle and read end (before per-base noise).