A multi-tier data mining workflow to analyze the age related shift from diglycosylated- to tetra-glycosylated-FSH secretion by the anterior pituitary
Mahesh Visvanathan, Shanker Rao, Vijay Anand, Gerald H Lushington, Michael Netzer, Christian Baumgartner, George R Bousfield; A Multi-Tier Data Mining Workflow to Analyze the Age Related Shift from Diglycosylated- to Tetra-glycosylated-FSH Secretion by the Anterior Pituitary; BCB '10 Proceedings of the First ACM International Conference on Bioinformatics and Computational Biology Pages 508-514; ISBN: 978-1-4503-0438-2 doi>10.1145/1854776.1854874
FSH is a glycoprotein hormone secreted as two major glycosylation variants by the anterior pituitary, which regulates reproduction in adults. As FSH consists of two functionally significant glycoforms, differentially expressed genes related to FSH biosynthesis in the anterior pituitary can help us to understand implications of changes in their relative abundance at the genomic level. Mapping these kinds of biomarker genes and their corresponding pathways is a key technology for studying the elaboration of FSH variants that affect the reproductive system. In this paper we use a multiple tier data mining work flow to identify FSH biosynthesis-related genes in the anterior pituitary. Our methodology combines different filterbased feature selection mechanisms like Linear Regression (LR), Z-Score statistics and the Biomarker Identifier (BMI). Consequently, we identified differentially expressed genes in response to the synthetic estrogen, diethylstilbestrol (DES), treatment in male rats. As a next step, we performed pathway analysis to identify the most relevant metabolic pathways associated with a set of identified genes in a pathway. Finally, we applied Mutual Information (MI) to calculate the measure of association between differentially expressed genes and several biosynthetic and signaling pathways of interest.