Upload README.md with huggingface_hub

README.md

@@ -1,7 +1,6 @@
 ---
 license:
 - mit
-- other
 license_link: https://github.com/Biohub/esm/blob/main/THIRD_PARTY_NOTICE.md
 language: en
 tags:
@@ -27,7 +26,7 @@ ESMFold2 is capable of either single-sequence or MSA conditioned structure predi
 
 To run this model with the Biohub Platform API, visit the [Biohub Platform](https://biohub.ai/).
 
-Read more about ESMFold2 in our paper [here](https://biohub.ai/papers/esm_protein.pdf).
+Read more about ESMFold2 in our paper [here](https://www.biorxiv.org/content/10.64898/2026.06.03.729735).
 
 ## Model Variants
 
@@ -42,7 +41,7 @@ ESMfold2 was evaluated against state-of-the-art single-sequence and MSA-based st
 
 ![][image1]
 
-Refer to the [paper](https://biohub.ai/papers/esm_protein.pdf) for details on additional performance metrics.
+Refer to the [paper](https://www.biorxiv.org/content/10.64898/2026.06.03.729735) for details on additional performance metrics.
 
 ### Usage
 
@@ -164,6 +163,8 @@ result = client.fold_all_atom(spi, config=FoldingConfig(num_loops=3, num_samplin
 print(f"pLDDT mean: {float(result.plddt.mean()):.3f}, pTM: {float(result.ptm):.3f}, ipTM: {float(result.iptm):.3f}")
 ```
 
+For multi-chain complexes (e.g. antibody–antigen), you can supply paired MSAs to take advantage of inter-chain co-evolution. See the [ESMFold2 tutorial](https://github.com/biohub/esm/blob/main/cookbook/tutorials/esmfold2.ipynb) for details.
+
 ## Training Data
 
 ESMfold2 was trained on sequences from the Protein Data Bank (PDB) and the AlphaFold DB (AFDB).
@@ -216,7 +217,7 @@ Should you have any security or privacy issues or questions related to this mode
             and Pannu, Jassi and Bachas, Sharrol and Liu, Daniel S.
             and Sercu, Tom and Rives, Alexander},
   year   = {2026},
-  url    = {https://biohub.ai/papers/esm_protein.pdf},
+  url    = {https://www.biorxiv.org/content/10.64898/2026.06.03.729735},
   note   = {Preprint}
 }
 ```