Overview

Nabo is a flexible Python package that allows projections of cells from one population to another. Nabo works by setting one of the populations as a reference’ and then maps cells from other populations (‘targets’) onto it. Nabo provides data implicit methods of verifying mapping quality, this allows users to clearly infer similarities between sub-populations across samples.

Comparing two or more cell populations can be of interest due to multiple reasons:

• Identifying cells of origin
• Comparison across replicates
• Testing population proportions change across conditions

Why use Nabo:

• No assumptions with respect to nature of data. One can use raw counts or normalized data.
• Optimized for speed. Nabo calculates distances using Numba’s JIT compilers
• Data persistence. All data is saved in HDF5 format files and data can be reanalyzed from any step.
• Low memory footprint. Nabo uses out-of-core functions such that only a fraction of data is loaded into the memory at once. Hence, Nabo can easily be run on laptops with modest memory sizes.

Unique features of Nabo:

• Allows mapping multiple samples over the same reference dataset
• Visual and statistical comparison of different projections
• Uses graph-optimized hierarchical clustering approach to identify cell groups in the reference population
• Provides inbuilt methods to generate null expectation projections
• Compatible with multiple data formats.