Back

WrapEvoFS enables auditable feature compression with regret-constrained representative locking

Park, J.; Schmidt, E.; Chi, S.; Hassanabad, M. F.; Foster, L. J.; Zhong, H.

2026-10-04 bioinformatics
10.64898/2026.09.28.755112 bioRxiv
Show abstract

WrapEvoFS provides a deterministic, auditable rule for choosing one feature set from repeated stochastic searches. It retains candidates within a prespecified development-score gap, selects the most representative set by feature overlap, and records the decision before held-out evaluation. A corrected genetic objective preserves score ordering lost through zero truncation. In a fully nested four-class cancer analysis, all 30 locked signatures met the 0.01 tolerance, and 11 differed from the highest-scoring candidate; every changed selection increased mean feature overlap within the eligible pool. Paired differences in area under the receiver operating characteristic curve against recursive feature elimination with cross-validation all had 95% intervals crossing zero. No selector dominated matched comparisons on feature count and predictive performance, and one-time evaluation of frozen signatures at held-out centers yielded heterogeneous effects. WrapEvoFS standardizes selection within a fixed candidate bank under an empirical score-gap constraint, enabling reproducible reporting of performance, compression, and feature-agreement trade-offs without implying predictive superiority or stability under changes in participant composition.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

1
Nature Communications
○ 5641 papers in training set
Top 14%
12.8%
Similar papers in this journal
2
Cell Systems
○ 201 papers in training set
Top 0.2%
11.8%
Similar papers in this journal
3
Genome Medicine
○ 183 papers in training set
Top 0.1%
10.6%
Similar papers in this journal
4
npj Precision Oncology
○ 53 papers in training set
Top 0.3%
4.8%
Similar papers in this journal
5
PLOS ONE
● 5266 papers in training set
Top 30%
4.8%
Similar papers in this journal
6
Briefings in Bioinformatics
◐ 354 papers in training set
Top 3%
3.2%
7
Cancer Research
● 130 papers in training set
Top 1%
3.2%
50% of probability mass above
8
Nature Methods
○ 385 papers in training set
Top 3%
3.2%
9
Nucleic Acids Research
◐ 1281 papers in training set
Top 6%
3.2%
10
Nature Machine Intelligence
○ 70 papers in training set
Top 1%
2.1%
11
Genetics in Medicine
○ 78 papers in training set
Top 0.5%
2.1%
12
Genome Biology
○ 637 papers in training set
Top 5%
2.1%
13
Bioinformatics
◐ 1204 papers in training set
Top 6%
2.1%
14
JCO Clinical Cancer Informatics
● 22 papers in training set
Top 0.4%
1.7%
15
Scientific Reports
○ 3612 papers in training set
Top 57%
1.5%
16
npj Digital Medicine
○ 118 papers in training set
Top 2%
1.5%
17
Nature Biomedical Engineering
○ 47 papers in training set
Top 1.0%
1.4%
18
Nature Genetics
○ 286 papers in training set
Top 4%
1.3%
19
Communications Medicine
○ 113 papers in training set
Top 2%
1.3%
20
Science Advances
● 1243 papers in training set
Top 26%
1.1%
21
PLOS Computational Biology
● 1863 papers in training set
Top 17%
1.1%
22
Molecular Systems Biology
○ 162 papers in training set
Top 3%
1.0%
23
Nature Biotechnology
○ 172 papers in training set
Top 4%
1.0%
24
Cell Reports Medicine
○ 153 papers in training set
Top 4%
1.0%
25
Cancer Cell
○ 42 papers in training set
Top 1%
0.9%
26
npj Systems Biology and Applications
○ 125 papers in training set
Top 2%
0.8%
27
Cell Reports Methods
○ 165 papers in training set
Top 4%
0.8%
28
BMC Medical Genomics
○ 50 papers in training set
Top 1%
0.8%
29
Communications Biology
○ 993 papers in training set
Top 35%
0.6%
30
Cancer Discovery
● 66 papers in training set
Top 2%
0.6%

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.