ADAPT

ADAPT analyzes differential abundance for microbiome metagenomics data by pooling Tobit models and using non-differentially abundant taxa as a reference for hypothesis testing.

Key Features:

• Left-censored zero handling: ADAPT treats zero counts as left-censored observations to facilitate computation and enhance interpretation.
• Tobit models for log count ratios: ADAPT uses Tobit models to model log count ratios for differential abundance analysis.
• Reference taxa selection: ADAPT employs a theoretically justified method to select non-differentially abundant microbiome taxa as a reference for hypothesis testing.
• Phyloseq format compatibility: ADAPT accepts microbiome metagenomics data in phyloseq format.

Scientific Applications:

• 16S rRNA sequencing analysis: ADAPT analyzes 16S rRNA sequencing data, as demonstrated on saliva samples from infants in the COHRA2 study.
• Shotgun metagenomics sequencing analysis: ADAPT analyzes shotgun metagenomics sequencing data, as demonstrated on plaque samples from infants in the COHRA2 study.
• Oral microbiome and dental caries association studies: ADAPT identifies associations between the oral microbiome and early childhood dental caries.

Methodology:

ADAPT pools Tobit models for log count ratios with left-censored zero counts and selects non-differentially abundant taxa as reference for hypothesis testing.