New data analysis and mining approaches identify unique proteome and transcriptome markers of susceptibility to autoimmune diabetes

Ivan Gerling, Sudhir Singh, Nataliya I. Lenchik, Dana R. Marshall, Jian Wu

Research output: Contribution to journalArticle

48 Citations (Scopus)

Abstract

Non-obese diabetic (NOD) mice spontaneously develop autoimmunity to the insulin producing beta cells leading to insulin-dependent diabetes. In this study we developed and used new data analysis and mining approaches on combined proteome and transcriptome (molecular phenotype) data to define pathways affected by abnormalities in peripheral leukocytes of young NOD female mice. Cells were collected before mice show signs of autoimmunity (age, 2-4 weeks). We extracted both protein and RNA from NOD and C57BU/6 control mice to conduct both proteome analysis by two-dimensional gel electrophoresis and transcriptome analysis on Affymetrix expression arrays. We developed a new approach to analyze the two-dimensional gel proteome data that included two-way analysis of variance, cluster analysis, and principal component analysis. Lists of differentially expressed proteins and transcripts were subjected to pathway analysis using a commercial service. From the list of 24 proteins differentially expressed between strains we identified two highly significant and interconnected networks centered around oncogenes (Myc and Mycn) and apoptosis-related genes (Bcl2 and Casp3). The 273 genes with significant strain differences in RNA expression levels created six interconnected networks with a significant over-representation of genes related to cancer, cell cycle, and cell death. They contained many of the same genes found in the proteome networks (including Myc and Mycn). The combination of the eight, highly significant networks created one large network of 272 genes of which 82 had differential expression between strains either at the protein or the RNA level. We conclude that new proteome data analysis strategies and combined information from proteome and transcriptome can enhance the insights gained from either type of data alone. The overall systems biology of prediabetic NOD mice points toward abnormalities in regulation of the opposing processes of cell renewal and cell death even before there are any clear signatures of immune system activation.

Original languageEnglish (US)
Pages (from-to)293-305
Number of pages13
JournalMolecular and Cellular Proteomics
Volume5
Issue number2
DOIs
StatePublished - Feb 1 2006

Fingerprint

Data Mining
Proteome
Medical problems
Type 1 Diabetes Mellitus
Transcriptome
Data mining
Genes
Inbred NOD Mouse
Cell death
RNA
Autoimmunity
Analysis of Variance
Proteins
Cell Death
Gels
Insulin
myc Genes
Systems Biology
Gene Regulatory Networks
Immune system

All Science Journal Classification (ASJC) codes

  • Analytical Chemistry
  • Biochemistry
  • Molecular Biology

Cite this

New data analysis and mining approaches identify unique proteome and transcriptome markers of susceptibility to autoimmune diabetes. / Gerling, Ivan; Singh, Sudhir; Lenchik, Nataliya I.; Marshall, Dana R.; Wu, Jian.

In: Molecular and Cellular Proteomics, Vol. 5, No. 2, 01.02.2006, p. 293-305.

Research output: Contribution to journalArticle

@article{dc9ae88701734fd5942a2c36b30dec1d,
title = "New data analysis and mining approaches identify unique proteome and transcriptome markers of susceptibility to autoimmune diabetes",
abstract = "Non-obese diabetic (NOD) mice spontaneously develop autoimmunity to the insulin producing beta cells leading to insulin-dependent diabetes. In this study we developed and used new data analysis and mining approaches on combined proteome and transcriptome (molecular phenotype) data to define pathways affected by abnormalities in peripheral leukocytes of young NOD female mice. Cells were collected before mice show signs of autoimmunity (age, 2-4 weeks). We extracted both protein and RNA from NOD and C57BU/6 control mice to conduct both proteome analysis by two-dimensional gel electrophoresis and transcriptome analysis on Affymetrix expression arrays. We developed a new approach to analyze the two-dimensional gel proteome data that included two-way analysis of variance, cluster analysis, and principal component analysis. Lists of differentially expressed proteins and transcripts were subjected to pathway analysis using a commercial service. From the list of 24 proteins differentially expressed between strains we identified two highly significant and interconnected networks centered around oncogenes (Myc and Mycn) and apoptosis-related genes (Bcl2 and Casp3). The 273 genes with significant strain differences in RNA expression levels created six interconnected networks with a significant over-representation of genes related to cancer, cell cycle, and cell death. They contained many of the same genes found in the proteome networks (including Myc and Mycn). The combination of the eight, highly significant networks created one large network of 272 genes of which 82 had differential expression between strains either at the protein or the RNA level. We conclude that new proteome data analysis strategies and combined information from proteome and transcriptome can enhance the insights gained from either type of data alone. The overall systems biology of prediabetic NOD mice points toward abnormalities in regulation of the opposing processes of cell renewal and cell death even before there are any clear signatures of immune system activation.",
author = "Ivan Gerling and Sudhir Singh and Lenchik, {Nataliya I.} and Marshall, {Dana R.} and Jian Wu",
year = "2006",
month = "2",
day = "1",
doi = "10.1074/mcp.M500197-MCP200",
language = "English (US)",
volume = "5",
pages = "293--305",
journal = "Molecular and Cellular Proteomics",
issn = "1535-9476",
publisher = "American Society for Biochemistry and Molecular Biology Inc.",
number = "2",

}

TY - JOUR

T1 - New data analysis and mining approaches identify unique proteome and transcriptome markers of susceptibility to autoimmune diabetes

AU - Gerling, Ivan

AU - Singh, Sudhir

AU - Lenchik, Nataliya I.

AU - Marshall, Dana R.

AU - Wu, Jian

PY - 2006/2/1

Y1 - 2006/2/1

N2 - Non-obese diabetic (NOD) mice spontaneously develop autoimmunity to the insulin producing beta cells leading to insulin-dependent diabetes. In this study we developed and used new data analysis and mining approaches on combined proteome and transcriptome (molecular phenotype) data to define pathways affected by abnormalities in peripheral leukocytes of young NOD female mice. Cells were collected before mice show signs of autoimmunity (age, 2-4 weeks). We extracted both protein and RNA from NOD and C57BU/6 control mice to conduct both proteome analysis by two-dimensional gel electrophoresis and transcriptome analysis on Affymetrix expression arrays. We developed a new approach to analyze the two-dimensional gel proteome data that included two-way analysis of variance, cluster analysis, and principal component analysis. Lists of differentially expressed proteins and transcripts were subjected to pathway analysis using a commercial service. From the list of 24 proteins differentially expressed between strains we identified two highly significant and interconnected networks centered around oncogenes (Myc and Mycn) and apoptosis-related genes (Bcl2 and Casp3). The 273 genes with significant strain differences in RNA expression levels created six interconnected networks with a significant over-representation of genes related to cancer, cell cycle, and cell death. They contained many of the same genes found in the proteome networks (including Myc and Mycn). The combination of the eight, highly significant networks created one large network of 272 genes of which 82 had differential expression between strains either at the protein or the RNA level. We conclude that new proteome data analysis strategies and combined information from proteome and transcriptome can enhance the insights gained from either type of data alone. The overall systems biology of prediabetic NOD mice points toward abnormalities in regulation of the opposing processes of cell renewal and cell death even before there are any clear signatures of immune system activation.

AB - Non-obese diabetic (NOD) mice spontaneously develop autoimmunity to the insulin producing beta cells leading to insulin-dependent diabetes. In this study we developed and used new data analysis and mining approaches on combined proteome and transcriptome (molecular phenotype) data to define pathways affected by abnormalities in peripheral leukocytes of young NOD female mice. Cells were collected before mice show signs of autoimmunity (age, 2-4 weeks). We extracted both protein and RNA from NOD and C57BU/6 control mice to conduct both proteome analysis by two-dimensional gel electrophoresis and transcriptome analysis on Affymetrix expression arrays. We developed a new approach to analyze the two-dimensional gel proteome data that included two-way analysis of variance, cluster analysis, and principal component analysis. Lists of differentially expressed proteins and transcripts were subjected to pathway analysis using a commercial service. From the list of 24 proteins differentially expressed between strains we identified two highly significant and interconnected networks centered around oncogenes (Myc and Mycn) and apoptosis-related genes (Bcl2 and Casp3). The 273 genes with significant strain differences in RNA expression levels created six interconnected networks with a significant over-representation of genes related to cancer, cell cycle, and cell death. They contained many of the same genes found in the proteome networks (including Myc and Mycn). The combination of the eight, highly significant networks created one large network of 272 genes of which 82 had differential expression between strains either at the protein or the RNA level. We conclude that new proteome data analysis strategies and combined information from proteome and transcriptome can enhance the insights gained from either type of data alone. The overall systems biology of prediabetic NOD mice points toward abnormalities in regulation of the opposing processes of cell renewal and cell death even before there are any clear signatures of immune system activation.

UR - http://www.scopus.com/inward/record.url?scp=33644669952&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33644669952&partnerID=8YFLogxK

U2 - 10.1074/mcp.M500197-MCP200

DO - 10.1074/mcp.M500197-MCP200

M3 - Article

VL - 5

SP - 293

EP - 305

JO - Molecular and Cellular Proteomics

JF - Molecular and Cellular Proteomics

SN - 1535-9476

IS - 2

ER -