Comparative evaluation of DNA extraction protocols for neonatal gut microbiota profiling in a resource-limited setting

Accurate and reproducible gut microbiome profiling depends heavily on the DNA extraction method, particularly in low-biomass samples such as neonatal stool. In this study, we evaluated the performance of three commercially available DNA extraction kits; QIAamp Fast DNA Stool Mini, DNeasy PowerSoil Pro, and ZymoBIOMICS DNA Miniprep on neonatal stool samples collected in a resource-limited hospital in Kano, Nigeria. Samples were stored under various conditions (temperature and preservatives), and DNA was extracted and sequenced using Oxford Nanopore Technologies. DNA yield differed substantially across extraction kits and storage conditions. The bead-beating-based kits: PowerSoil and ZymoBIOMICS, consistently outperformed the QIAamp Fast DNA Stool Mini kit, which produced negligible yields across all conditions. Both bead-beating kits achieved the highest DNA concentrations when samples were processed fresh and without preservatives, while yields declined sharply after just one day of storage. Although overall DNA yields were similar between PowerSoil and ZymoBIOMICS at all time points, PowerSoil extracts produced longer sequencing reads and higher-quality assemblies. Specifically, PowerSoil-derived libraries generated higher read-level N50 values than those from ZymoBIOMICS at both Day 0 and Week 6. While these differences were not statistically significant, a moderate effect size at Day 0 (rank-biserial correlation = 0.60) suggests a potential advantage in genome assembly continuity. Additionally, PowerSoil had a shorter processing time, enhancing its suitability for long-read metagenomics workflows in resource-limited settings. We conclude that same-day processing using bead-beating-based extraction kits improves yield and may reduce bias in neonatal gut microbiome studies. These findings are especially relevant for low-resource settings, where equipment limitations and delayed sample processing can impact data quality and study scalability.