Update README.md

README.md

@@ -11,16 +11,16 @@ tags:
 
 # PXR Challenge Multi-method Cofolding
 
-This dataset contains matched ensemble cofolding predictions (OpenFold3, Boltz-2, and Chai1) for the nuclear hormone receptor hPXR and >600 ligands. 
+This dataset contains >40,000 cofolding predictions using three different methods (OpenFold3, Boltz2, Chai1) for the nuclear hormone receptor hPXR and with >600 ligands. 
 Ligands are derived from the OpenADMET PXR Challenge and contain known binders, known agonists, and inactives.  More on the ligand set can be found here: https://huggingface.co/spaces/openadmet/pxr-challenge
 
-Each dataset was run using default settings, except where noted, and 20-25 cofolding samples per system (5 samples * 4 or 5 seeds)
+Each cofolding method was run using default settings and untemplated, except where noted.  Each protein/ligand/method/parameter set generated 20-25 structures (5 samples * 4 or 5 random seeds)
 
 Deposited is the raw cofolding output with all confidence predictions and other data files. 
 
 ### Contents:
 - `input_data`: raw SMILES tables, PDB64 rerefined input structures, hPXR LBD sequence, and ColabFold MSA used for all cofolding runs.
-- `pdb64`: cofolding predictions for a 70-system rerefined PXR crystal set. Untemplated and templated with the protein crystall coordinates.
+- `pdb64`: cofolding predictions for a 70-system rerefined PXR crystal set. Untemplated and templated with the protein crystal coordinates.
 - `structure_challenge`: cofolding predictions and inputs for 184 structure-challenge ligands.
 - `activity_challenge`: cofolding predictions for 513 activity-challenge ligands.