Struct SBModel 

pub struct SBModel { /* private fields */ }

The sequence-specific bias Markov model.

Implementations

impl SBModel

pub fn new() -> Self

pub fn dump(&self) -> &[f64]

The flattened transition table (probs[pos * ROWS + idx]), for dumping to the aux bias files. Linear counts before normalize, conditional log-probabilities after.

pub fn add_context(&mut self, context: &[u8], rev_comp: bool, weight: f64)

Accumulate one observed context with the given weight.

pub fn normalize(&mut self)

Convert accumulated counts into conditional log-probabilities. Idempotent guard: a model can only be normalized once.

Two steps: divide each group of four counts (the four possible bases given one preceding context) by their total, turning counts into conditional probabilities; then take logs so evaluate_log can add instead of multiply.

pub fn evaluate_log(&self, context: &[u8], rev_comp: bool) -> f64

Log-probability the (normalized) model assigns to a context.

A sum of nine per-position log-probabilities, which is the product of the nine conditional probabilities — the chain rule for a Markov model.

pub fn is_trained(&self) -> bool

pub fn combine_counts(&mut self, other: &SBModel)

Add another (un-normalized) model’s counts into this one. Both must be pre-normalization; used to merge per-thread observed models.

Trait Implementations

impl Clone for SBModel

fn clone(&self) -> SBModel

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for SBModel

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for SBModel

fn default() -> Self

Returns the “default value” for a type.