DeepSeek as the paradigm shift in rare disease diagnosis – the power of a fully automated genetic variant classification system
Bibliographic record
Abstract
Abstract Large language models (LLMs) have been extensively tested for incorporating into medical applications in recent years, yet their potential in clinical genetics, particularly in diagnosing rare diseases, remains underexplored. Recent advancements in LLMs have improved their reasoning capabilities and transparency, facilitating significant enhancements in clinical workflow designs. The open-sourced DeepSeek model also serves as a cost-effective alternative of top-ranked proprietary reasoning LLMs such as o3-mini-high for genome projects and hospitals that have specific needs in data security. In this study, we developed a framework that fully automates genetic variant classification according to the American College of Medical Genetics and Genomics (ACMG) and the Association for Molecular Pathology (AMP) guidelines and Clinical Genome Resource (ClinGen) recommendations. Two state-of-the art LLMs, DeepkSeek-R1 and o3-mini-high were tested for their performance in variant classification. We demonstrated that through careful prompt engineering and creation of ACMG-rule specific knowledgebases, DeepSeek-R1 outperformed o3-mini-high and achieved high sensitivity and 100% specificity in interpreting ACMG rules that require understanding literature-based evidence. Further testing using 150 variants curated by ClinGen experts, DeepSeek-R1 demonstrated performance on par with human curators. Finally, we showed the framework can be also used for reanalysis using 150 ClinVar variants with conflicting interpretations. Our study provided the first LLM framework capable of fully automated variant classification in the diagnosis of genetic diseases and variant reanalysis.
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 | 0.004 | 0.009 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.002 | 0.003 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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".