A Machine Learning Approach to Identifying Causal Monogenic Variants in Inflammatory Bowel Disease
Bibliographic record
Abstract
Background and AimsDiagnosis of monogenic disease is increasingly important for patient care and personalizing therapy. However, the current process is nonstandardized, expensive, and time consuming. There is currently no accepted strategy to help identify disease-causing variants in monogenic inflammatory bowel disease (IBD). The aim of the study is to develop a prioritization strategy for monogenic IBD variant discovery through detailed analysis of a whole-exome sequencing (WES) data set.MethodsAll consenting pediatric patients with IBD presenting to our tertiary care hospital during the study period were enrolled and underwent WES (n = 1005). Available family members also underwent WES. Variants were analyzed en masse using the GEMINI framework and were further annotated using data from dbNSFP, Combined Annotation Dependent Depletion, and gnomAD. Known disease-causing variants (n = 36) were used as positive controls. Machine learning algorithms were optimized and then compared to assist with identifying monogenic IBD case characteristics.ResultsInitial gene-level analysis identified 11 genes not previously linked to IBD that could potentially harbor IBD-causing variants. Machine learning algorithms identified 4 primary variant characteristics (Combined Annotation Dependent Depletion score, dbNSFP score, relationship with a known immunodeficiency gene, and alternate allele frequency), and optimal threshold values for each were determined to assist with identifying monogenic IBD variants. Based on these characteristics, an automated variant prioritization pipeline was then created that filters and prioritizes variants from >100,000 variants per patient down to a mean of 15. This pipeline is available online for all to use.ConclusionLeveraging a large WES data set, we demonstrate a statistically rigorous strategy for prioritization of variants for monogenic IBD diagnosis. Diagnosis of monogenic disease is increasingly important for patient care and personalizing therapy. However, the current process is nonstandardized, expensive, and time consuming. There is currently no accepted strategy to help identify disease-causing variants in monogenic inflammatory bowel disease (IBD). The aim of the study is to develop a prioritization strategy for monogenic IBD variant discovery through detailed analysis of a whole-exome sequencing (WES) data set. All consenting pediatric patients with IBD presenting to our tertiary care hospital during the study period were enrolled and underwent WES (n = 1005). Available family members also underwent WES. Variants were analyzed en masse using the GEMINI framework and were further annotated using data from dbNSFP, Combined Annotation Dependent Depletion, and gnomAD. Known disease-causing variants (n = 36) were used as positive controls. Machine learning algorithms were optimized and then compared to assist with identifying monogenic IBD case characteristics. Initial gene-level analysis identified 11 genes not previously linked to IBD that could potentially harbor IBD-causing variants. Machine learning algorithms identified 4 primary variant characteristics (Combined Annotation Dependent Depletion score, dbNSFP score, relationship with a known immunodeficiency gene, and alternate allele frequency), and optimal threshold values for each were determined to assist with identifying monogenic IBD variants. Based on these characteristics, an automated variant prioritization pipeline was then created that filters and prioritizes variants from >100,000 variants per patient down to a mean of 15. This pipeline is available online for all to use. Leveraging a large WES data set, we demonstrate a statistically rigorous strategy for prioritization of variants for monogenic IBD diagnosis.
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 | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| 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".