BAscVI

This repository accompanies the paper:

“Adversarial learning enables unbiased organism-wide cross-species alignment of single-cell RNA data at scale”
arXiv preprint

---

Repository Overview

This repo contains three Python modules that can be installed separately depending on your use case:

• pai
  API for inference of cell-type embeddings and cell-type labels from scRNA-seq datasets.
  Supports .h5ad formats (downloadable from cellxgene).

• pai_soma_data
  A wrapper around TileDB for exploring the scREF atlas, used in our public notebooks.

• ml_benchmarking
  Code to run and evaluate ML models on the scREF and scREF-mu benchmarks.

---

Installation

Clone the repository:

git clone https://github.com/PhenomicAI/bascvi.git
cd bascvi

pip install pai - recomended

pip install -U phenomic-ai  

Install pai from local

TODO: update github version to 1.11

pip install -e src/pai

Verify:

from pai.utils.option_choices import tissue_organ_option_choices
from pai.embed import PaiEmbeddings

Install pai_soma_data

pip install -e src/pai_soma_data

Verify:

from pai_soma_data import pai_soma_data

Install ml_benchmarking

pip install -e src/ml_benchmarking

Verify:

import ml_benchmarking.bascvi as bascvi

You can install any combination of these modules depending on your needs.

---

Example Usage

from pai.utils.option_choices import tissue_organ_option_choices
from pai.embed import PaiEmbeddings
from pai_soma_data import pai_soma_data
import ml_benchmarking.bascvi as bascvi

---

License

Everything is released under an MIT license. Please feel free to use it, but cite us as we have cited others.

---