Graph Representation Learning for the Prediction of Medication Usage in the UK Biobank Based on Pharmacogenetic Variants
Bibliographic record
Abstract
Ineffective treatment and side effects are associated with high burdens for the patient and society. We investigated the application of graph representation learning (GRL) for predicting medication usage based on individual genetic data in the United Kingdom Biobank (UKBB). A graph convolutional network (GCN) was used to integrate interconnected biomedical entities in the form of a knowledge graph as part of a machine learning (ML) prediction model. Data from The Pharmacogenomics Knowledgebase (PharmGKB) was used to construct a biomedical knowledge graph. Individual genetic data (n = 485,754) from the UKBB was obtained and preprocessed to match with pharmacogenetic variants in the PharmGKB. Self-reported medication usage labels were obtained from UKBB data field 20003. We hypothesize that pharmacogenetic variants can predict the impact of medications on individuals. We assume that an individual using a medication on a regular basis experiences a net benefit (vs. side-effects) from the medication. ML models were trained to predict medication usage for 264 medications. The GCN model significantly outperformed both a baseline logistic regression model (p-value: 1.53 × 10−9) and a deep neural network model (p-value: 8.68 × 10−8). The GCN model also significantly outperformed a GCN model trained using a random graph (GCN-random) (p-value: 5.44 × 10−9). A consistent trend of medications with higher sample sizes having better performance was observed, and for several medications, a high relative rank of the medication (among multiple medications) was associated with greater than 2-fold higher odds of usage of the medication. In conclusion, a graph-based ML approach could be useful in advancing precision medicine by prioritizing medications that a patient may need based on their genetic data. However, further research is needed to improve the quality and quantity of genetic data and to validate our approach using more reliable medication labels.
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 | 0.002 | 0.012 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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".