pavm595 commited on
Commit
52fec32
·
verified ·
1 Parent(s): a85dd1c

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +6 -2
README.md CHANGED
@@ -31,10 +31,14 @@ MolE learns task-independent molecular representations of chemicals via Graph Is
31
 
32
  ## Long description
33
 
34
- MolE integrates molecular graph-based representation learning with gradient-boosted decision trees for predicting antimicrobial potential. The approach involves:
 
 
 
 
35
  1. **Representation learning:** A graph neural network (GINet) trained on 100,000 randomly sampled compounds to derive
36
  molecular embeddings from SMILES strings.
37
- 2. **Prediction:** These embeddings are used as input to an **XGBoost** model that predicts antimicrobial activity scores across 40 bacterial strains, based on data from *Maier et al., 2018*. The model was developed by **Roberto Olayo Alarcon et al.**. Further information is available in the [paper](https://www.nature.com/articles/s41467-025-58804-4).
38
 
39
  ## Metadata
40
 
 
31
 
32
  ## Long description
33
 
34
+ **MolE** is a non-contrastive self-supervised Graph Neural Network (GNN) framework that leverages unlabeled chemical structures to learn task-independent molecular representations. By combining a pre-trained MolE representation with experimentally validated compound-bacteria activity data, the project builds an antimicrobial prediction model that re-discovers recently reported growth-inhibitory compounds
35
+ that are structurally distinct from current antibiotics. Using the model as a compound prioritization strategy, three human-targeted drugs are identified and experimentally confirm as growth-inhibitors of Staphylococcus aureus, highlighting MolE’s potential to accelerate the discovery of new antibiotics.
36
+
37
+
38
+ <!-- MolE integrates molecular graph-based representation learning with gradient-boosted decision trees for predicting antimicrobial potential. The approach involves:
39
  1. **Representation learning:** A graph neural network (GINet) trained on 100,000 randomly sampled compounds to derive
40
  molecular embeddings from SMILES strings.
41
+ 2. **Prediction:** These embeddings are used as input to an **XGBoost** model that predicts antimicrobial activity scores across 40 bacterial strains, based on data from *Maier et al., 2018*. The model was developed by **Roberto Olayo Alarcon et al.**. Further information is available in the [paper](https://www.nature.com/articles/s41467-025-58804-4). -->
42
 
43
  ## Metadata
44