MAT-classifier: A memory-efficient pipeline for accurate genus-level profiling from ancient metagenomic data

Abstract With advances in sequencing technology, the prospect of studying microbes from ancient samples to reconstruct past environments and host–microbe interactions has been growing rapidly. However, the field remains constrained by computational challenges and accuracy problems due to the difficulty of validating truly ancient microbes within noisy datasets dominated by modern contaminants. Existing pipelines often demand substantial memory resources, limiting their use to researchers with access to advanced computational systems. Here, we present MAT-classifier, a pipeline for genus-level profiling of ancient taxa designed to increase accuracy while substantially reducing computational requirements. Using simulated bacterial datasets, we showed that the MAT-classifier achieves more accurate detection of ancient taxa with substantially lower memory usage and shorter runtime than the existing counterpart, the aMeta pipeline. Validation on deeply sequenced ancient metagenomic data further confirmed its low memory footprint and practical utility. Overall, the MAT-classifier provides a reliable, efficient, and accessible alternative to current pipelines, lowering technical barriers to enable broader adoption of ancient microbiome research. Competing Interest Statement The authors have declared no competing interest.