Data Analysis in Experimental Biomedical Research

This thesis covers two non-related topics in experimental biomedical research: data analysis in thrombin generation experiments (collaboration with Novo Nordisk A/S), and analysis of images and physiological signals in the context of neurovascular signalling and blood flow regulation in the brain (collaboration with University of Copenhagen).
The ongoing progress in experimental methods of thrombin generation allowed to introduce ready-to-use commercial assays for thrombin measurement. On the other hand, commercial assays use “black box” data analysis which makes it nearly impossible for researches to critically assess and compare obtained results. We reverse engineered the data analysis performed by CAT, a de facto standard assay in the field. This revealed a number of possibilities to improve its methods of data analysis. We found that experimental calibration data is described well with textbook Michaelis-Menten kinetics of the basic enzymatic reaction. We also found that a simple phenomenological model inspired by a cascade of basic enzymatic reactions describes thrombin generation data. Finally, we proved that CAT greatly overestimates several key physiological parameters that describe thrombin generation. Our work is a step towards standardization in the field of thrombin generation based on transparent and well-motivated flexible methods of data analysis.
One of the recent surprising discoveries about pericytes was their active role in the regulation of the blood flow. However, this discovery was critically acclaimed in literature, so their true regulatory function has only started to emerge. We were able to bridge the two opposing points of view and found that pericytes are likely to be the local centers of capillary blood flow regulation. In addition, we investigated the role of octapeptide angiotensin two (ANGII) in the mechanism of hypertension on neurovascular signaling and regulation in the brain. We found that values of resting and increase in cerebral blood flow dropped upon administration of ANGII in a concentration-dependent manner, in agreement with previous findings. We also observed local constrictions of vessels in up to second order, which proves that ANGII changes their contractile tone by directly affecting smooth muscle cells as well as a subset of capillary pericytes. Here we developed tools for noise-robust extraction of vessel diameter traces from experimental two-photon microscopy images that could serve the broad community of biologists and neuroscientists. More importantly, we demonstrated on a number of examples how both careful data management and rigorous statistical analysis are crucial for deriving solid conclusions about the behavior of complex biological systems.