Microarray Gene Expression for Predicting Histo-Clinical Variables in Kidney Transplant Biopsies.
Bibliographic record
Abstract
Linear discriminant analysis (LDA) was used to test the ability of microarray gene expression to predict histo-clinical variables in for-cause kidney transplant biopsies. Predictions in 703 biopsies were evaluated using the area under the curve (AUC) in test sets, using repeated 10-fold cross-validation. The standard classifier (SC) used the top 20 genes in each training set. The time-adjusted classifier (TAC) used the top 19 genes after controlling for time post-transplant, and time as the 20th predictor. Time alone (not using LDA) was used as a control since it is freely available at the time of biopsy, and clearly related to at least some lesions. Results are summarized in Table 1. Column 1 shows the split-point for predictions, e.g. g>0 vs g=0. AUCs (based on the same split-points) for the 3 models are shown next, followed by the number of genes significant at fdr=0.05, and the top 3 genes in the TAC model. Thousands of genes were significant for all variables except cv. In general, TAC produced the highest AUCs, and these were significantly higher (p < 0.001) than time alone for all variables except cv and ah. These were also the variables with the fewest and most weakly associated genes. TAC was significantly better than SC for g, cg, ci, ct, mm, ah, and DSA. Predictions from simple gene set scores (not using LDA) were inferior to TAC for all variables except i and t (results not shown). When searching for gene associations, time post-transplant should be taken into account to avoid spurious correlations with time itself. For most variables, this leads to better predictors and more biologically informative gene sets.Table: No Caption available.DISCLOSURES:Halloran, P.: Other, Astellas, lecturing, Novartis, lecturing, One Lambda, lecturing.
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 | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| 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".