ON THE COMPLEMENTARITY OF THE CONSENSUS-BASED DISORDER PREDICTION
Bibliographic record
Abstract
Intrinsic disorder in proteins plays important roles in transcriptional regulation, translation, and cellular signal transduction. The experimental annotation of the disorder lags behind the rapidly accumulating number of known protein chains, which motivates the development of computational predictors of disorder. Some of these methods address predictions of certain types/flavors of the disorder and recent years show that consensus-based predictors provide a viable way to improve predictive performance. However, the selection of the base predictors in a given consensus is usually performed in an ad-hock manner, based on their availability and with a premise that more is better. We perform first-of-its-kind investigation that analyzes complementarity among a dozen recent predictors to identify characteristics of (future) predictors that would lead to further consensus-based improvements in the predictive quality. The complementarity of a given set of three base predictors is expressed by the differences in their predictions when compared with each other and with their majority vote consensus. We propose a regression-based model that quantifies/predicts quality of the majority-vote consensus of a given triplet of predictors based on their individual predictive performance and their complementarity measured at the residue and the disorder segment levels. Our model shows that improved performance is associated with higher (lower) similarity between the three base predictors at the residue (segment) level and to their consensus prediction at the segment (residue) level. We also show that better consensuses utilize higher quality base methods. We use our model to predict the best-performing consensus on an independent test dataset and our empirical evaluation shows that this consensus outperforms individual methods and other consensus-based predictors based on the area under the ROC curve measure. Our study provides insights that could lead to the development of a new generation of the consensus-based disorder predictors.
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 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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, 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".