Work

KIRP: Detecting allele-specific expression with k-mer analysis of RNAseq data

Year: 2026

Type: article

Abstract: Epigenetic effects such as genomic imprinting have been shown to play a role in the endosperm of developing seeds of the model organism, Arabidopsis thaliana. Experiments on other plant species must overcome challenges of outcrossing, heterozygosity, and lower quality references. To expand the study of allelic expression to other Arabidopsis species, we present a new protocol that has low reliance on reference data. KIRP uses k-mer analysis of paired RNAseq read sequences from three individuals: two parents and their cross. It builds k-mer sets specific to each parent and deletes their intersection. It assigns cross reads to a parent, where possible, by set comparison. It bins reads by gene using a map to the nearest available annotated reference. It applies standard differential expression statistics to the binned read counts. KIRP was designed to avoid the pitfalls of map bias that arise when parental genomes differ by level of heterozygosity and similarity to the (single-haplotype) reference. We present KIRP results on crosses of two wild ecotypes of the species Arabidopsis arenosa. We replicated experiments on four crosses in each direction (i.e. swapping the pollen donor and acceptor), and we filtered for genes expressed in endosperm. Compared to a reference-based analysis of the same data, the KIRP results indicated greater resilience to the confounding factors of heterozygosity and map bias.

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.22283993

Authors: Miller, Jason Rafe, Hornslien, Karina, Grini, Paul

Institutions: Hood College, University of Oslo

Language: English

Topic: Genomics and Phylogenetic Studies

Subfield: Molecular Biology

Field: Biochemistry, Genetics and Molecular Biology

Domain: Life Sciences

SDG: Zero hunger

Open Access status: green

Funder: Norges Forskningsråd