Update README.md

README.md

@@ -23,10 +23,14 @@ where
 - `model_loc`: the location of the model weights `.torch` file.
 - `batch_size`: the per GPU batch size. For the 33 layer model, on a 80GB GPU, you should use 25. For a 4 layer model on the same GPU, you can use 100.
 
+For a sample output on the 10k pbmc dataset, run
+```
+python eval_single_anndata.py
+```
+All necessary model files will be downloaded automatically.
 
-This script also makes use of additional files, which are described in the code documentation. These are downloaded automatically unless already present in the working directory.
 
-For a sample output on the 10k pbmc dataset, run `eval_single_anndata.py` with default parameters.
+**Note**: This script makes use of additional files, which are described in the code documentation. These are downloaded automatically unless already present in the working directory. The script defaults to the pretrained 4-layer model. For running the pretrained 33-layer model from the paper, please download using this [link](https://figshare.com/articles/dataset/Universal_Cell_Embedding_Model_Files/24320806?file=43423236) and set `--nlayers 33`.
 
 ## Output
 
@@ -39,3 +43,18 @@ outputs from `eval_single_anndata.py` are stored in the `dir` directory.
 ## Data
 
 You can download processed datasets used in the papere [here](https://drive.google.com/drive/folders/1f63fh0ykgEhCrkd_EVvIootBw7LYDVI7?usp=drive_link)
+
+## Citing
+
+If you find our paper and code useful, please consider citing the [preprint](https://www.biorxiv.org/content/10.1101/2023.11.28.568918v1):
+
+```
+@article{rosen2023universal,
+  title={Universal Cell Embeddings: A Foundation Model for Cell Biology},
+  author={Rosen, Yanay and Roohani, Yusuf and Agrawal, Ayush and Samotorcan, Leon and Consortium, Tabula Sapiens and Quake, Stephen R and Leskovec, Jure},
+  journal={bioRxiv},
+  pages={2023--11},
+  year={2023},
+  publisher={Cold Spring Harbor Laboratory}
+}
+```