Using molecular embeddings in QSAR modeling: does it make a difference?
Bibliographic record
Abstract
With the consolidation of deep learning in drug discovery, several novel algorithms for learning molecular representations have been proposed. Despite the interest of the community in developing new methods for learning molecular embeddings and their theoretical benefits, comparing molecular embeddings with each other and with traditional representations is not straightforward, which in turn hinders the process of choosing a suitable representation for Quantitative Structure-Activity Relationship (QSAR) modeling. A reason behind this issue is the difficulty of conducting a fair and thorough comparison of the different existing embedding approaches, which requires numerous experiments on various datasets and training scenarios. To close this gap, we reviewed the literature on methods for molecular embeddings and reproduced three unsupervised and two supervised molecular embedding techniques recently proposed in the literature. We compared these five methods concerning their performance in QSAR scenarios using different classification and regression datasets. We also compared these representations to traditional molecular representations, namely molecular descriptors and fingerprints. As opposed to the expected outcome, our experimental setup consisting of over $25 000$ trained models and statistical tests revealed that the predictive performance using molecular embeddings did not significantly surpass that of traditional representations. Although supervised embeddings yielded competitive results compared with those using traditional molecular representations, unsupervised embeddings tended to perform worse than traditional representations. Our results highlight the need for conducting a careful comparison and analysis of the different embedding techniques prior to using them in drug design tasks and motivate a discussion about the potential of molecular embeddings in computer-aided drug design.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.004 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".