Split-Transformer Impute (STI): A Transformer Framework for Genotype Imputation
Bibliographic record
Abstract
Abstract Motivation Despite recent advances in sequencing technologies, genome-scale datasets continue to have missing bases and genomic segments. Such incomplete datasets can undermine downstream analyses, such as disease risk prediction and association studies. Consequently, the imputation of missing information is a common pre-processing step for which many methodologies have been developed. However, the imputation of genotypes of certain genomic regions and variants, including large structural variants, remains a challenging problem. Results Here, we present a transformer-based deep learning framework, called a split-transformer impute (STI) model, for accurate genome-scale genotype imputation. Empowered by the attention-based transformer model, STI can be trained for any collection of genomes automatically using self-supervision. STI handles multi-allelic genotypes naturally, unlike other models that need special treatments. STI models automatically learned genome-wide patterns of linkage disequilibrium (LD), evidenced by much higher imputation accuracy in high LD regions. Also, STI models trained through sporadic masking for self-supervision performed well in imputing systematically missing information. Our imputation results on the human 1000 Genomes Project show that STI can achieve high imputation accuracy, comparable to the state-of-the-art genotype imputation methods, with the additional capability to impute multi-allelic structural variants and other types of genetic variants. Moreover, STI showed excellent performance without needing any special presuppositions about the patterns in the underlying data when applied to a collection of yeast genomes, pointing to easy adaptability and application of STI to impute missing genotypes in any species.
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 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.004 | 0.002 |
| Research integrity | 0.002 | 0.004 |
| Insufficient payload (model declined to judge) | 0.006 | 0.002 |
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".