Analysis of covariance patterns in gene expression data and FT-IR spectra

Abstract

The aim of this study was to detect and interpret correlation patterns in several large data matrices from the same biological system using Partial Least Squares Regression (PLSR) in order to get information on the system under investigation. To do this, DNA microarray data and Fourier Transform Infrared (FT-IR) spectra from a designed study where Campylobacter jejuni was exposed to environmental stress conditions, were used. The experimental design included variation in atmospheric conditions, temperature and time. PLSR was first used to analyse each of the two data types separately in order to explore the effect of the experimental parameters on the data. The results showed that both the gene expression and FT-IR spectra were affected by the variations in atmosphere, temperature and time, but that the effect was different for the two types of data. When the DNA microarray data and FT-IR spectra were linked together by PLSR, covariation due to temperature was seen. Both specific genes and ranges in the FT-IR spectra that were connected to the variation in temperature were detected. Some of these are possibly connected to properties of the cell wall of the bacteria. The results in this study show the potential of PLSR for investigation of covariance structures in biological data. By doing this, valuable information about the biological system can be detected and interpreted. It was also shown that the use of FT-IR spectroscopy provided important information about the stress responses in the bacteria, information that was not detected from the DNA microarray data.