Datasets:
Sara Hantgan
commited on
Commit
·
b27f1b9
1
Parent(s):
a029564
Add final project files
Browse files- .gitattributes +1 -0
- KiDatabase.csv +3 -0
- README.md +33 -0
- curated_ki_database.csv +3 -0
- fingerprints_with_ki.csv +3 -0
- rf_model.pkl +3 -0
- test_predictions.csv +3 -0
- train_model.ipynb +0 -0
.gitattributes
CHANGED
|
@@ -57,3 +57,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 57 |
# Video files - compressed
|
| 58 |
*.mp4 filter=lfs diff=lfs merge=lfs -text
|
| 59 |
*.webm filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 57 |
# Video files - compressed
|
| 58 |
*.mp4 filter=lfs diff=lfs merge=lfs -text
|
| 59 |
*.webm filter=lfs diff=lfs merge=lfs -text
|
| 60 |
+
*.csv filter=lfs diff=lfs merge=lfs -text
|
KiDatabase.csv
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:93a15227e00534bed1d6e6bc777af679b66eb8180c797fbb18248df4602aeefe
|
| 3 |
+
size 23869637
|
README.md
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Serotonin Receptor (5-HT) Binding Affinity Prediction Dataset
|
| 2 |
+
|
| 3 |
+
This dataset was curated from the PDSP Ki Database to support training machine learning models that predict binding affinity (Ki in nM) of ligands to serotonin (5-HT) receptors.
|
| 4 |
+
|
| 5 |
+
## Files Included
|
| 6 |
+
|
| 7 |
+
- `curated_ki_database.csv`: Cleaned Ki dataset filtered for 5-HT targets
|
| 8 |
+
- `fingerprints_with_ki.csv`: Molecules represented as 2048-bit Morgan (ECFP4) fingerprints
|
| 9 |
+
- `rf_model.pkl`: Trained scikit-learn RandomForestRegressor model
|
| 10 |
+
- `test_predictions.csv`: Predictions vs. actual Ki values on the test set
|
| 11 |
+
- `train_model.ipynb`: Full Jupyter notebook with training code and evaluation
|
| 12 |
+
|
| 13 |
+
## Modeling Approach
|
| 14 |
+
|
| 15 |
+
- Molecules were standardized using [MolVS](https://molvs.readthedocs.io).
|
| 16 |
+
- SMILES strings were converted to 2048-bit Morgan fingerprints (ECFP4) using RDKit.
|
| 17 |
+
- A Random Forest Regressor was trained using scikit-learn.
|
| 18 |
+
- The model was evaluated on a held-out 20% test set.
|
| 19 |
+
|
| 20 |
+
### 📈 Model Performance
|
| 21 |
+
|
| 22 |
+
- **R² Score**: `your_R2_here`
|
| 23 |
+
- **RMSE**: `your_RMSE_here` nM
|
| 24 |
+
|
| 25 |
+
## Source
|
| 26 |
+
|
| 27 |
+
- PDSP Ki Database: https://pdsp.unc.edu/databases/kidb.php
|
| 28 |
+
|
| 29 |
+
## Author
|
| 30 |
+
|
| 31 |
+
Sara Hantgan
|
| 32 |
+
University of Michigan | BIOINF 595 Final Project
|
| 33 |
+
Winter 2025
|
curated_ki_database.csv
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:293026aca70bc7325ad6ff95f82e3c19bf674176a1fc259f46e057ade9391c57
|
| 3 |
+
size 1203983
|
fingerprints_with_ki.csv
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0677f24c0a3d8dbf8823866cf35c9362aecc72d1e689847c9977fe2f1f9ee458
|
| 3 |
+
size 46742625
|
rf_model.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:69f4eb943d00eb4c4229d77c30e6c5b2e6a7f1769002242548c9f70efad840f2
|
| 3 |
+
size 12905809
|
test_predictions.csv
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:48a388b910f74d8670de8c8d360c1ff90f1a6b523ecdda78246438e29305a266
|
| 3 |
+
size 57428
|
train_model.ipynb
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|