Full text
| |
| Author(s): |
Victor Sande Vasconcelos
Total Authors: 1
|
| Document type: | Master's Dissertation |
| Press: | Campinas, SP. |
| Institution: | Universidade Estadual de Campinas (UNICAMP). Instituto de Biologia |
| Defense date: | 2019-02-21 |
| Examining board members: |
José Andrés Yunes;
Marcelo Falsarella Carazzolle;
Samara Flamini Kiihl
|
| Advisor: | José Andrés Yunes |
| Abstract | |
Gene expression analysis has been widely used to understand biological phenomena and to discriminate Acute Lymphocytic Leukemia (ALL) subtypes. Gene expression data are usually obtained based on the same RNA mass per sample. Differences in cellular volume are disregarded in regular statistical analysis, leading to results that may not be consistent with reality. In this work, we measure cell size and performed a different normalization strategy to investigate how individual gene expression behaves along cellular volume increase. Previous work showed that gene expression has a linear trend when the cell volume increases but each gene has its own linear pattern (Padovan-Merhar et al., 2015). Because of that, gene expression normalization based on RNA mass and/or global expression average results in a volume-dependent deviation in the expression value that is not consistent with actual expression per cell. The aim of this study is to propose a method for gene expression normalization according to cell volume, obtaining, in this way, a volume-dependent correction in the expression values. The work was done using Affymetrix microarray data from 91 cases of ALL (91 precursor B-cell ALL), for which we also have an estimate of the cellular size by flow cytometry and bone marrow smears. Statistical analysis and calculations were done using the R software. Some packages as "Affy" and "genefilter" Packages (Bioconductor) were used to obtain the expression expectation and variability, as a cell volume function, for 32,321 genes. Principal Component Analysis has been used in every volume spectrum fragment. The ALL samples used in this work showed a 5.5 fold-change variation between the smallest and biggest average cell volume. Comparison of smaller versus bigger cells, within each molecular subgroup of ALL, revealed around 1,000 genes that have a fold-change greater than 2 and low standard deviation throughout the volume spectrum. Genes with higher volume-dependent fold-changes belong to diverse pathway and functions. However, cell cycle genes and histones appear frequently. Furthermore, unsupervised PCA analyses, over the volume spectrum, showed that important genes for the discrimination between ALL subtypes change as a function of cell volume. The expression of a single gene in different ALL samples shows diverse volume-dependent trends. Differential gene expression analysis using current normalization strategies resulted in lost of important information (AU) | |
| FAPESP's process: | 17/02301-9 - Normalization of acute lymphoblastic leukemia gene expression data using the cellular volume |
| Grantee: | Victor Sande Vasconcelos |
| Support Opportunities: | Scholarships in Brazil - Master |