Machine Learning Model for Repurposing Drugs to Target Viral Diseases
Bibliographic record
Abstract
With recent events, such as the Covid-19 pandemic, it is increasingly important to develop strategies to combat viral diseases. Due to technological advancements, computer-aided drug design and machine learning (ML)-based hit identification strategies have gained popularity. Applying these techniques to identify novel scaffolds and/or repurpose existing therapeutics for viral diseases is a promising approach. As an avenue to improve existing classification models for antiviral applications, this thesis aimed to make improvements to non-binding data selection within these models. We created a classification model using molecular fingerprints to assess the performance of machine learning predictions when the model is trained using randomly selected and rationally selected non-binding datasets. Our analyses revealed that machine learning predictions can be improved using a rational selection approach. We further used this approach and trained three machine learning models based on XGBoost, Random Forest, and Support Vector Machine to predict potential inhibitors for the SARS-CoV2 main protease (Mpro) enzyme. Probability-ranked hits from the combined model were further analyzed using classical structure-based methods. The binding modes and affinities of the hits were identified using AutoDock Vina, and molecular dynamics simulations-enabled MM-GBSA calculations. The top hits identified from this multi-step screening approach revealed potential candidates that show improved affinity and stability than existing non-covalent Mpro inhibitors. Thus, our approach and the model could be useful for screening large ligand libraries.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 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 source (direct Gemma or distilled Codex), 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".