Multiplex protein data and how to handle it
Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
Objective: To nd which methods and techniques that work well when working with multiplex proteindata. A short analysis of a dataset to show the methods is done at the end of the thesis.
Material and methods: A study containing 171 individuals is used to test if there are any dierencesin protein values based on obesity, sugar level and type of obesity(5 subgroups). To investigate this OlinksProseek multiplex is used to generate 92 dierent protein values for each individual in the dataset. MainlyWelch's t-test and ANOVA are used as the comparison method for multiplex protein data. Since multipletests are performed the p-values are adjusted to avoid multiple type I error. Dierent ways to visualize thedata are also presented.
Results: When comparing the obese and non-obese groups 25 proteins are shown to be signicantly dierent(p-value <0:05). Low and high sugar have no proteins that are signicantly dierent(p-value <0:05) betweenthe two groups and for the dierent types of obesity 18 proteins are signicant dierent(p-value <0:05).
Conclusion: Both the tests for obesity and types of obesity show dierences between the groups while thesugar level do not have any proteins that signicantly dier between the groups. However when performinga multivariate techniques of low and high sugar the two can be separated indicating some dierences inprotein values.
Place, publisher, year, edition, pages
2016. , 30 p.
Probability Theory and Statistics
IdentifiersURN: urn:nbn:se:uu:diva-297390OAI: oai:DiVA.org:uu-297390DiVA: diva2:941684
Master Programme in Statistics