fast.adonis: a computationally efficient non-parametric multivariate analysis of microbiome data for large-scale...

Area of Science:

• Environmental science
• Ecology
• Bioinformatics

Background:

• Nonparametric multivariate analysis is crucial for identifying variables associated with dissimilarity matrices.
• Existing methods like adonis/adonis2 are computationally intensive for large studies and complex sampling designs.
• Adjusting for sampling weights is necessary for unbiased estimates in complex surveys.

Purpose of the Study:

• To develop a computationally efficient R function for large-scale nonparametric multivariate analysis.
• To address the limitations of existing software in handling large datasets and complex sampling designs.
• To provide an accurate and fast alternative for analyzing ecological and environmental data.

Main Methods:

• Implementation of an R function named fast.adonis.
• Integration of sampling weights algebraically to account for complex survey designs.
• Leveraging efficient algorithms to reduce computational time and memory requirements.

Main Results:

• fast.adonis achieves results consistent with adonis/adonis2 but with significantly faster computation.
• The function efficiently handles large datasets and complex sampling designs without excessive memory usage.
• Accurate and unbiased estimates are obtained even with the inclusion of sampling weights.

Conclusions:

• fast.adonis offers a computationally efficient solution for large-scale nonparametric multivariate analysis.
• The function is suitable for studies with complex sampling designs requiring the use of sampling weights.
• The R package is publicly available, facilitating its adoption in ecological and environmental research.