Discuss A legume-enriched diet improves metabolic health in prediabetes mediated through gut microbiome: a randomized controlled trial
LinDA (Linear Models for Differential Abundance)[edit source]
This study used LinDA (Linear Models for Differential Abundance) as the primary statistical method to identify taxa and genes that were differentially abundant between the intervention and control groups over time.
This is explicitly stated in the text associated with Figure 3F:
"Linear Models for Differential Abundance (LinDA) analysis identified 30 species (using a cut-off q(Treatment×Time) < 0.15, q(Time) < 0.05) and 15 genera..."
LinDA was applied on centered log-ratio (CLR)-transformed microbiome data and incorporated both treatment group and time as fixed effects. Multiple hypothesis testing was corrected using the Benjamini-Hochberg procedure, which aligns with common standards in microbiome differential abundance testing.
Additional confirmation appears in the Methods section, which further specifies:
- Use of LinDA for species, genus, and gene-level analyses.
- Use of LMMs (linear mixed models) for clinical and continuous biomarker data, but not for identifying DA taxa.
