Update README.md
Browse files
README.md
CHANGED
@@ -35,6 +35,15 @@ Note that targets have missing values. Algorithms should be evaluated only on pr
|
|
35 |
| Recommended split | scaffold |
|
36 |
| Recommended metric | AUPRC, AUROC |
|
37 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
38 |
## References
|
39 |
<a id="1">[1]</a>
|
40 |
Ramsundar, Bharath, et al.
|
|
|
35 |
| Recommended split | scaffold |
|
36 |
| Recommended metric | AUPRC, AUROC |
|
37 |
|
38 |
+
**Warning:** in newer RDKit vesions, 2 molecules from the original dataset are not read correctly due to disallowed
|
39 |
+
hypervalent states of some atoms (see [release notes](https://github.com/rdkit/rdkit/releases/tag/Release_2024_09_1)).
|
40 |
+
This version of the PCBA dataset contains manual fixes for those molecules, removing additional hydrogens, e.g. `[AlH3] -> [Al]`.
|
41 |
+
In OGB scaffold split, used for benchmarking, both molecules are in the training set. Applied mapping is:
|
42 |
+
```
|
43 |
+
"CC(=O)O[AlH3]OC(C)=O" -> "CC(=O)O[Al]OC(C)=O"
|
44 |
+
"CCCCO[AlH3](OCCCC)OCCCC" -> "CCCCO[Al](OCCCC)OCCCC"
|
45 |
+
```
|
46 |
+
|
47 |
## References
|
48 |
<a id="1">[1]</a>
|
49 |
Ramsundar, Bharath, et al.
|