Genetic Analysis Workshop 18 single-nucleotide variant prioritization based on protein impact, sequence conservation, and gene annotation
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Grouping variants based on gene mapping can augment the power of rare variant association tests. Weighting or sorting variants based on their expected functional impact can provide additional benefit. We defined groups of prioritized variants based on systematic annotation of Genetic Analysis Workshop 18 (GAW18) single-nucleotide variants; we focused on variants detected by whole genome sequencing, specifically on the high-quality subset presented in the genotype files. First, we divided variants between coding and noncoding. Coding variants are fewer than 1% of the total and are more likely to have a biological effect than noncoding variants. Coding variants were further stratified into protein changing and protein damaging groups based on the effect on protein amino acid sequence. In particular, missense variants predicted to be damaging, splice-site alterations, and stop gains were assigned to the protein damaging category. Impact of noncoding variants is more difficult to predict. We decided to rely uniquely on conservation: we combined (a) the mammalian phastCons Conserved Element and (b) the PhyloP score, which identify conserved intervals and the single-nucleotide position, respectively. This reduced the noncoding variants to a number comparable to coding variants. Finally, using gene structure definition from the widely used RefSeq database, we mapped variants to genes to support association tests that require collapsing rare variants to genes. Companion GAW18 papers used these variant priority groups and gene mapping; one of these paper specifically found evidence of stronger association signal for protein damaging variants.
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.
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 | 0.000 |
| 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 it