API: The MerTools submodule

A simple mer count struct.

MerCount is a simple struct that binds a mer value to a count of the number of times it has been observed. This type, (sorted) vectors of them, and some additional utility methods, form the basic building blocks of the higher-level mer counting functionality of the MerTools sub-module.

A type for storing a frequency histogram of MerCounts, also sometimes referred to as a Kmer spectra.

Shorthand for MerCount{DNAMer{K}}

Shorthand for MerCount{RNAMer{K}}

Get the mer from a MerCount.

Get the count from a MerCount.

Get the count from a MerCount, and convert it to type R.

collapse_into_counts(mers::Vector{M}) where {M<:AbstractMer}

Build a vector of sorted MerCounts from a Vector of a mer type.

This is a basic kernel function used for any higher level and more complex kmer counting procedures.

collapse_into_counts!(result::Vector{MerCount{M}}, mers::Vector{M}) where {M<:AbstractMer}

Build a vector of sorted MerCounts from a Vector of a mer type.

This is a basic kernel function used for any higher level and more complex kmer counting procedures.

This is like collapse_into_counts, except it's first argument is a result vector that is cleared and filled with the result.

merge_into!(a::Vector{MerCount{M}}, b::Vector{MerCount{M}}) where {M<:AbstractMer}

Merge the MerCounts from vector b into the vector a.

build_freq_list(::Type{M}, sbuf::SequenceBuffer{PairedReads}, range::UnitRange{Int}) where {M<:AbstractMer}

Build a sorted list (vector) of kmer counts (MerFreq), serially and in memory.

This function is a serial and in memory MerFreq list builder that can build a kmer count from a PairedReads datastore on its own (if you have memory and time), but it is also intended to be composed into other multi-process or multi-threaded kmer counting strategies.

This method estimates roughly how many kmers will be generated by the reads specified by range in the dataset. It then pre-allocates an array to contain them. It then collects the kmers, sorts, them, and then collapses them into a list of counts sorted by the kmer.

build_freq_list(::Type{M}, sbuf::SequenceBuffer{PairedReads}, range::UnitRange{Int}, chunk_size::Int) where {M<:AbstractMer}

Build a sorted list (vector) of kmer counts (MerFreq), serially and in memory.

This function is a serial and in memory MerFreq list builder that can build a kmer count from a PairedReads datastore on it own (if you have memory and time), but it is also intended to be composed into other multi-process or multi-threaded kmer counting strategies.

This method pre-allocates space for chunk_size kmers, and iterates over kmers in the reads in the dataset specified by range until the buffer is filled. The mers are then collapsed into a list of counts, sorted by the kmer. This list is then merged into another output list. This process repeats for many chunks of kmers, building up the output list.

This method is useful for situations where you don't want (or have the space) to allocate a buffer to collect all the kmers in the dataset all in one go.

unsafe_collapse_into_counts!(result::Vector{MerCount{M}}, mers::Vector{M}) where {M<:AbstractMer}

unsafe_merge_into!(a::Vector{MerCount{M}}, b::Vector{MerCount{M}}) where {M<:AbstractMer}

Merge the MerCounts from vector b into the vector a.