Work

Comparing k-mer Methods and Language Model Embeddings Under Cross-Species Evaluation

Year: 2026

Type: preprint

Abstract: This narrative review compares k-mer-based representations and biological language-model embeddings for biological sequence analysis. It focuses on homology leakage, pretraining overlap, benchmark design, and cross-species evaluation, and proposes practical recommendations for more reliable model assessment.

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.21711164

Authors: Subedi, Prasiddha, Sahni, Shivansh

Language: English

Topic: Machine Learning in Bioinformatics

Subfield: Molecular Biology

Field: Biochemistry, Genetics and Molecular Biology

Domain: Life Sciences

Open Access status: green

OpenAlex ID: W7171820573

Cites:

Cited by:

Related to: