Reproducing the manual annotation of multiple sequence alignments using a SVM classifier
Bibliographic record
Abstract
MOTIVATION: Aligning protein sequences with the best possible accuracy requires sophisticated algorithms. Since the optimal alignment is not guaranteed to be the correct one, it is expected that even the best alignment will contain sites that do not respect the assumption of positional homology. Because formulating rules to identify these sites is difficult, it is common practice to manually remove them. Although considered necessary in some cases, manual editing is time consuming and not reproducible. We present here an automated editing method based on the classification of 'valid' and 'invalid' sites. RESULTS: A support vector machine (SVM) classifier is trained to reproduce the decisions made during manual editing with an accuracy of 95.0%. This implies that manual editing can be made reproducible and applied to large-scale analyses. We further demonstrate that it is possible to retrain/extend the training of the classifier by providing examples of multiple sequence alignment (MSA) annotation. Near optimal training can be achieved with only 1000 annotated sites, or roughly three samples of protein sequence alignments. AVAILABILITY: This method is implemented in the software MANUEL, licensed under the GPL. A web-based application for single and batch job is available at http://fester.cs.dal.ca/manuel. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.
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 | 0.016 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.004 |
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".