Exploring Multivariate Data

Document Type

Book Chapter

Publication Title

Indian Statistical Institute Series

Abstract

Moving from one gene to two genes naturally leads to the idea of the existence and consideration of multiple genes together. Not only genes, but this may be true for phenotypes also. If we suspect that more than two phenotypes need to be jointly considered while studying any genetic association, the same issue arises. If we have expression data for more than two genes, we might study them separately. We can use a similar idea for two or more phenotypes occurring together. Dealing with a single gene or one phenotype at a time seems okay, but there is a high possibility that we might lose the inherent correlation structure among multiple genes or multivariate phenotypes. On the other hand, a statistical way of treating them together seems more realistic; this might retain the correlation structures among genes and is more informative to decipher the genetic architecture. But how do we look at such types of data? Depending on our outlook, questions, and the problem under consideration, there are different ways and various statistical concepts and methods to deal with it. The idea of correlation among the variables is probably the major reason to consider variables together, i.e. from a multivariate perspective. While discussing correlation, we have studied correlation and regression in the context of two variables occurring together. Similarly in the multivariate set up we might ask the question as how and to what extent a variable is related to other variables.

First Page

213

Last Page

249

DOI

10.1007/978-981-99-3220-7_8

Publication Date

1-1-2023

This document is currently not available here.

Share

COinS