Abstract
A central challenge in medicine is translating from observational understanding to mechanistic understanding, where some observations are recognized as causes for the others. This can lead not only to new treatments and understanding, but also to recognition of novel phenotypes. Here, we apply a collection of mathematical techniques (empirical dynamics), which infer mechanistic networks in a model-free manner from longitudinal data, to hematopoiesis. Our study consists of three subjects with markers for cyclic thrombocytopenia, in which multiple cells and proteins undergo abnormal oscillations. One subject has atypical markers and may represent a rare phenotype. Our analyses support this contention, and also lend new evidence to a theory for the cause of this disorder. Simulations of an intervention yield encouraging results, even when applied to patient data outside our three subjects. These successes suggest that this blueprint has broader applicability in understanding and treating complex disorders.
Original language | English (US) |
---|---|
Article number | 100138 |
Journal | Patterns |
Volume | 1 |
Issue number | 9 |
DOIs | |
State | Published - Dec 11 2020 |
Externally published | Yes |
Keywords
- DSML 2: Proof-of-Concept: Data science output has been formulated, implemented, and tested for one domain/problem
- blood disorders
- causal inference
- complex disorders
- immunology
ASJC Scopus subject areas
- General Decision Sciences