kMate

kMate estimates allele frequencies for SNPs, short indels and large structural variants from pooled sequencing of a population descended from a known set of founder haplotypes.

It is alignment-free: reads are never mapped. Working in k-mer space is why structural variants are handled exactly like SNPs.

---

How it works, in two steps

Step 1: put your founder haplotypes into k-mer space

Turn your founder haplotypes' pangenome into two matrices, once per founder set and reused for every pool.

A founder haplotype is one continuous sequence, one allele per variant. An inbred line is one haplotype; a phased outbred individual is two. See Haplotypes and windows.

                                    ┌─ K_pa : founder × k-mer
   founders' pangenome (a VCF) ─────┤
                                    └─ V_pa : founder × variant

matrix	what it says	what it is for
K_pa	for each k-mer, which founder haplotypes contain it	the evidence: what the reads are compared against
V_pa	for each variant, which founder haplotypes carry the ALT allele	the translation: turns haplotype proportions into allele frequencies

A third file, var_called, records where each founder haplotype actually had a call, so that missing data is excluded rather than silently counted as reference.

→ Building a panel

Step 2: run a pool against them

Count the k-mers in the pooled reads and solve for the mixture that produced them: each count is modelled as Poisson with mean λ · Σ h_f · K_pa[f,k], and the founder-haplotype mixture h is fitted by expectation-maximisation.

Projecting h through V_pa gives each variant's frequency: the summed proportion of haplotypes carrying it. One pass covers every SNP, indel and SV.

   pooled reads ──▶ k-mer counts ──▶ EM ──▶ h (haplotype mixture) ──▶ × V_pa ──▶ allele frequencies

→ Running kMate

---

Start here

Installation	install kMate and check it works
Input files	what kMate needs, and what each file must satisfy
Haplotypes and windows	what a founder haplotype is, and whether to run per chromosome or per window
Building a panel	step 1: build the matrices from your founder haplotypes
Running kMate	step 2: estimate frequencies for a pool
Output	the result table, and how to read it
Command reference	every option, with its default
Troubleshooting	common errors and what they mean

---

The short version

# install
mamba create -n kmate -c conda-forge -c bioconda kmate "python<3.13"
mamba activate kmate
kmate selftest                      # must print PASS

Step 1: build the panel from your founders' VCF (once):

kmate build-index    --vcf graph.vcf.gz --ref REF.fa --out index/ours -k 31 --haploid   # graph VCF

kmate build-kmer-pa  --kmers index/ours_Chr1_kmers.tsv.gz \
                     --vcf panel.vcf.gz --ref REF.fa --chrom Chr1 \
                     --out kmer_pa/kmer_pa_Chr1 \
                     --treat-missing-as-n --filter-production

kmate build-var-pa   --vcf panel.vcf.gz --chrom Chr1 --out var_pa/var_pa_Chr1

Step 2: run each pool against them (per sample):

kmate run \
    --kmer-pa-prefix kmer_pa/kmer_pa \
    --var-pa-prefix  var_pa/var_pa \
    --reads R1.fq.gz R2.fq.gz \
    --sample MYPOOL --out MYPOOL_Chr1.tsv \
    --chroms Chr1 --unit chrom

Step 1 requires a haploid, biallelic, sequence-resolved VCF. Building a panel covers how to get one, including from raw assemblies.

---

Source, issues, citation

Code and the detailed technical docs: https://github.com/Tatianabellagio/kMate

kMate was developed for the GrENE-Net outdoor evolution experiment in Arabidopsis thaliana. It is not yet published as a standalone method. If you want to use it, or collaborate, get in touch: Tatiana Bellagio ([email protected]).

The panel-building path builds on tools from HPRC and eblerjana; see Building a panel.