Inferring gene flow between populations using statistical methods
Bibliographic record
Abstract
Gene flow is the transfer of genetic material from one population to another. It is very common and important in the description of the genetic history of a population. Gene flow is hidden in the genome as migrated segments of different genetic material. Inferring the gene flow based on sequenced genomes is challenging. I present my work on estimating gene flow using three different statistical models. With two sequences from different populations, the model Isolation Mi- gration CoalHMM can determine the amount of gene flow between the pop- ulations after their initial split. It takes all possible migrated segments into account using the powerful HMM algorithms. In collaboration I have build new CoalHMM’s incorporating several pairs of sequences to infer direction and variation in the gene flow. I show that estimating direction and variation jointly is too hard. I show that estimating direction is possible with some uncertainty. I apply the methods to a dataset of extinct and extant elephants and show that there is extensive gene flow. Instead of considering all possible segments with an HMM, I examine the potential of only considering the most likely segments with particle filtering. I present challenges and advantageous choices when implementing a particle filter for this problem. The second statistical model infers gene flow from a covariance matrix be- tween several populations. Some relations between entries in the covariance matrix can only be explained by gene flow. In collaboration I have developed a method that fits the best phylogeny with gene flow events for an observed co- variance matrix. The method uses MCMC. A phylogeny with gene flow events is called an admixture graph and the method is called AdmixtureBayes. I show that AdmixtureBayes has a smaller error than the most popular admixture graph estimators on simulated data. AdmixtureBayes produces a posterior sample of admixture graphs and I demonstrate the possibilities with such a sample on a real dataset of Native American genomes. The last statistical model infers very recent gene flow by classifying hybrid individuals. The genome of a hybrid individual has big segments of alleles originating from different populations. Using the allele frequencies from those populations, it is possible to infer the segments. In collaboration, I implemented ImmediateAncestry which infers the segments and the most likely hybrid type with an HMM. I show that the classifier has good accuracy on simulated data. The classifier is not robust for a real dataset of chimpanzees, so I discuss reasons and remedies.
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 | 0.005 | 0.019 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.005 | 0.002 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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".