I have clinical data for 8 subjects with samples acquired overtime. The number and type of data collected is very variable. Since we are evaluating the efficacy of our treatment, overtime each individual reacted differently. I would like to do a PCA analysis to reduce the variables of interest and determine if patients with similar outcome group together depending of the variables we measured.
I used Graphpad Prism 9.0 and my problem is that I dont know how to introduce the data for the program to analyze the data overtime. I have been able to group the patients for each time point but I would like to group them also overtime. Is it that even possible? If so, how can I do it?
Interesting question! I'm not experienced in PCA but assisted on a very different project to assess and compare huge electrophysiology datasets between treatments and over time. Maybe you need some sort of clustering algorithm to compare each patient to every other at each timepoint, for weighted importance of each variable (i.e. some variables you might consider more important than others, in the comparison), and then over time - which we represented as arrows between each timepoint - Article Visualization of incrementally learned projection trajectori...
. Good luck to find a solution - it might mean looking for something outside of Prism too.
- Badges
- Science topic
- Similar topics
- Engineering
- Materials Engineering
- Powders