miloR

miloR performs single-cell differential abundance testing by modelling cell states as representative neighbourhoods on a nearest neighbour graph and conducting hypothesis testing with negative binomial generalized linear models or negative binomial generalized linear mixed models.

Key Features:

• Neighbourhood-based cell state modelling: Cell states are represented as neighbourhoods on a nearest neighbour graph rather than discrete clusters.
• Flexible hypothesis testing: Supports both negative binomial generalized linear models and negative binomial generalized linear mixed models for statistical testing.

Scientific Applications:

• Single-cell differential abundance analysis: Detects changes in cell state abundance between experimental conditions in single-cell datasets.

Methodology:

Cell states are modelled as representative neighbourhoods on a nearest neighbour graph. Hypothesis testing is performed using either a negative binomial generalized linear model or negative binomial generalized linear mixed model.