Privacy Preserving Genomic Data Imputation using Autoencoders
Bibliographic record
Abstract
Next-generation sequencing technologies have significantly increased the availability of genomic databases. These databases can be used for numerous applications in genome-wide association studies, such as disease prediction and precision medicine. Missing values in genotype data diminish the database quality. Genotype imputation is an essential low-cost tool in genomics that statistically infers missing genotype variants from an experimentally observed set of variants. Owing to the computationally intensive nature of the problem, genotype imputation is often outsourced to an external service provider. However, sharing genomic data as a plaintext raises privacy concerns and leaks sensitive information. Existing privacy-preserving approaches perform genotype imputation using linear classification models, which require training a separate classifier for each variant using previously labeled data. Self-supervised deep learning models, such as autoencoders, have recently become popular because of their ability to model complex patterns in genomic datasets and achieve significantly high accuracy. However, deep learning-based genotype imputation under a privacy-preserving setting remains largely unexplored. In this work, we propose a novel adaptation of an autoencoder-based genotype imputation model that preserves the privacy of sensitive genomic data. To our knowledge, ours is the first work to do so. Genomic data privacy is preserved using fully homomorphic encryption (FHE). FHE schemes enable Efficient computations over encrypted data but suffer from noise growth as the number of computations increases. To overcome the issue of noise growth due to computationally complex deep learning, we use neural network quantization, which considerably reduces the network size while achieving high accuracy, as demonstrated in our results. We present all the necessary parameters to perform deep learning-based genotype imputation in the privacy-preserving setting.
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 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| 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".