Abstract B102: Using subclonal dynamics to detect and quantify fitness advantages of resistant subclones in tumors under treatment
Bibliographic record
Abstract
Abstract Introduction: In a tumor undergoing evolution under treatment, two forces drive changes in the frequencies of genetically distinct subpopulations (subclones): drift and natural selection. To identify the presence of a subclone with a fitness advantage in a tumor undergoing treatment (i.e. a drug-resistant subclone), it is important to distinguish between these two forces. In this work, we use simulations of subclonal dynamics to propose a method of determining and quantifying the relative fitness of a subclone, given a change in its population fraction. Methods: To simulate subclonal dynamics, a Gillespie algorithm was used to model the stochastic birth and death events of a tumor with six subpopulations of initially equal size. We simulated subclonal dynamics under two conditions – one in which all subclones were drifting (“drift simulations”) and the other in which one subclone was resistant to treatment (“selection simulations”). In drift simulations, the growth and death rates of all subpopulations were set the same, while in selection simulations, the growth rate of only one subpopulation was increased. In each type of simulation, the distribution of the size of the largest subclone was recorded over time. From these distributions, we defined a time-dependent threshold of subpopulation size, for determining if a subpopulation had a significant fitness advantage. Selecting various thresholds of subpopulation size led to receiver operating characteristic (ROC) curves showing the accuracy of this method in identifying a subclone with a known fitness value. Furthermore, by assuming a normal distribution of a priori fitness advantages, a Bayesian estimator was formulated to predict the most probable fitness value. Results: In our simulations of subclonal dynamics, we found that in selection simulations the fitter subclone tended to quickly and robustly sweep to dominance in the tumor population (“fixate”). In drift simulations as well, one subclone would typically fixate but only after a longer period of time. Consistent with these observations, we found that at intermediate times the size distributions of the largest subclone fraction for drift versus selection simulations diverge. From this, we were able to choose thresholds in subpopulation size for accurately assessing the existence of a fitness advantage. For a subclone with a 15% growth advantage, this method could determine the significance of its relative fitness with greater than 90% sensitivity and specificity. In addition, we used this method to quantify a subclone’s fitness advantage as a roughly linear relationship to the increase in its subpopulation size over a given time. Conclusion: In this work, we demonstrate a method for inferring the presence of natural selection and the fitness advantage of a resistant subclone from changes in subclonal frequencies over time. Applications of this method to Next-Generation Sequencing data may allow the early identification of aggressive and/or resistant subclones in cancers. Citation Format: Michael Salazar, Andrew Chen, Madison Stoddard, T. Ryan Gregory, Arijit Chakravarty. Using subclonal dynamics to detect and quantify fitness advantages of resistant subclones in tumors under treatment [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference on Molecular Targets and Cancer Therapeutics; 2025 Oct 22-26; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2025;24(10 Suppl):Abstract nr B102.
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".