MAI

MAI implements a two-step mechanism-aware approach to imputing missing data in metabolomics datasets.

Key Features:

• Missing Mechanism Classification: Uses a random forest classifier to classify missing values as either Missing Completely at Random/Missing At Random (MCAR/MAR) or Missing Not At Random (MNAR).
• MCAR/MAR Imputation: Applies Bayesian Principal Component Analysis (BPCA), Multiple Imputation No-Skip K-Nearest Neighbors (Multi_nsKNN), or Random Forest for imputing MCAR/MAR-classified missing values.
• MNAR Imputation: Applies nsKNN or a single imputation approach for metabolites with left-censoring for imputing MNAR-classified missing values.

Scientific Applications:

• Metabolomics Data Imputation: Addresses missing data in metabolomics studies by distinguishing and appropriately handling different missing data mechanisms.

Methodology:

MAI uses a two-step approach: Step 1 employs a random forest classifier to classify missing values as MCAR/MAR or MNAR; Step 2 imputes missing values using mechanism-appropriate algorithms.