scverse
Integrate E-Distance Calculation by scverse pertpy into rapids_singlecell
Is your feature request related to a problem? Please describe. The e-distance (energy distance) is a powerful statistical measure used to quantify the similarity or dissimilarity between distributions. Performing e-distance calculations using the scverse pertpy library for large datasets can be quite time-consuming. As genome-wide perturb-seq technologies continue to mature, the need to analyze these large perturb-seq datasets becomes increasingly relevant. Integrating this functionality into rapids_singlecell would allow users to leverage this metric for fast computation in various single-cell analysis tasks, such as comparing perturbation effects, clustering, and trajectory inference.
Describe the solution you'd like Integrate E-Distance calculation by scverse pertpy into rapids_singlecell
Is there a CPU based implementation Please check out following notebooks by pertpy:
- https://github.com/scverse/pertpy-tutorials/blob/2bb242871bedb95834e100ba2cb5b96d82be6152/distances.ipynb
- https://github.com/scverse/pertpy-tutorials/blob/2bb242871bedb95834e100ba2cb5b96d82be6152/distance_tests.ipynb
