Detecting the “invisible fraction” bias in resurrection experiments
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
The resurrection approach is a powerful tool for estimating phenotypic evolution in response to global change. Ancestral generations, revived from dormant propagules, are grown side by side with descendent generations in the same environment. Phenotypic differences between the generations can be attributed to genetic change over time. Project Baseline was established to capitalize on this potential in flowering plants. Project participants collected, froze, and stored seed from 10 or more natural populations of 61 North American plant species. These will be made available in the future for resurrection experiments. One problem with this approach can arise if nonrandom mortality during storage biases the estimate of ancestral mean phenotype, which in turn would bias the estimate of evolutionary change. This bias-known as the "invisible fraction" problem-can arise if seed traits that affect survival during storage and revival are genetically correlated to postemergence traits of interest. The bias is trivial if seed survival is high. Here, I show that with low seed survival, bias can be either trivial or catastrophic. Serious bias arises when (i) most seeds deaths are selective with regard to the seed traits, and (ii) the genetic correlations between the seed and postemergence traits are strong. An invisible fraction bias can be diagnosed in seed collections that are family structured. A correlation between the family mean survival rate and the family mean of a focal postemergence trait indicates that seed mortality was not random with respect to genes affecting the focal trait, biasing the sample mean. Fortunately, family structure was incorporated into the sampling scheme for the Project Baseline collection, which will allow bias detection. New and developing statistical procedures that can incorporate genealogical information into the analysis of resurrection experiments may enable bias correction.
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 | 0.001 | 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