Create README_mcpp-dataset.md

README_mcpp-dataset.md

@@ -0,0 +1,23 @@
+Unified all-atom molecule generation with neural fields - MCPP Dataset
+
+Starting from 641 protein–MCP complexes obtained from the RCSB PDB, we employed a “mutate-then-relax” approach to curate a dataset comprising 
+of 186,685 MCP–protein complexes. (mcpp_dataset.tar.gz). Our strategy consists of randomly mutating the MCPs at 1 to 8 different sites, 
+using a list of 213 distinct amino acids. We relaxing them using fast-relax in Rosetta, which involves iterative cycles of side-chain 
+packing and all-atom minimization. Then, we selected the best mutated complexes based on the lowest interface scores. The source dataset comprises 
+lengths ranging from 4 to 25 amino acids with an average of 10. 78% of the MCPs contain one or more non-canonical amino acids, i.e. any amino acid 
+that is neither L-canonical nor D-canonical. We split the dataset into train (train_data.pt), test (test_data.pt) and validation (val_data.pt) 
+subsets using a clustering approach creating a test set consisting of 100 protein pockets.
+
+
+
+After downloading the dataset (mcpp_dataset.tar.gz), open the folder using:
+
+tar -xvzf mcpp_dataset.tar.gz
+
+The preprocessed split data from the mcpp_dataset can be found in :
+test set: test_data.pt
+train set: train_data.pt
+validation set: val_data.pt
+
+To generate MCP samples with Funcbind, copy the [split]_data.pt files into mcpp_dataset.
+