NCSA Melanoma Project


The Melanoma Project aimed to provide the technological infrastructure to support 21st century genomic research. Its target audience is researchers who study the response of human DNA to exposure to genotoxins, such as radiation from prolonged exposure to the sun. The aim of these researchers is to create a biologically faithful model of the human systems of response to DNA damage that accounts for all the complex interactions among cell cycle checkpoints and DNA repair following exposure to carcinogenic agents.

RENCI collaborated with Dr. William Kaufmann and his lab in the Lineberger Cancer Center at the University of North Carolina at Chapel Hill to develop tools and software to integrate massive amounts of protein-protein interaction data found in biological databases (i.e., BIND, REACTOME, HPRD, CPATH) with gene expression data generated for this project and visualization tools. These tools create a method for studying system-level response to DNA damage that is based on integrated genomic data. RENCI and the Kaufmann lab also collaborated to devise computer models of cell division and DNA damage repair mechanisms.

Research Findings

To maintain genomic stability in normal human cells, cell cycle checkpoints are activated when DNA damage occurs due to internal or external causes. Mutations in DNA damage responsive genes may reduce the effectiveness of checkpoint functions designed keep cells functioning normally and this can lead to the development of cancer. To elucidate the patterns of gene regulation associated with cell cycle checkpoint functions, changes in global gene expression were determined using Agilent Human 22k arrays after treating normal human diploid fibroblast lines—cells that help to maintain normal structural framework—with four DNA-damaging agents: ionizing radiation (IR), 254 nm ultraviolet radiation (UV), doxorubicin (Dox) and cadmium chloride (Cd).

To determine the roles of ATM and p53 in gene regulation in response to DNA damage, similar studies were performed with ATM- or p53-deficient fibroblast lines after radiation treatment. Quantitative flow cytometric analyses of cell cycle compartments (G1, S, G2 and M) showed that IR, UV, Dox and Cd, at doses inhibiting clonal expansion by 40 percent, caused genotoxin-specific changes in cell cycle progression and checkpoint function. Growth arrest in G1 and G2 was associated with repression of large numbers of cell cycle-regulated and growth-associated genes as determined by gene ontology analysis. The genotoxins also produced unique time-dependent gene expression. For example, UV induced unique time-dependent patterns of change in a set of 44 genes at 6 h post irradiation that was not induced by IR. Gene ontology analysis of this UV-specific pattern identified RNA transcription as the biological process controlled by the set of co-regulated genes. As expected, treatment with IR caused significantly less change in gene expression in ATM- and p53-deficient cell lines. Less induction of DNA-damage responsive genes soon after IR treatment and less repression of cell cycle-regulated genes at a later time were associated with the defective cell cycle checkpoint functions in these cell lines.


National Science Foundation under Grant No. CA SCI 0525308 and CSA SCI 0525308 through the National Center for Supercomputing Applications, University of Illinois-Urbana/Champaign.

Project Leaders

  • William Kaufmann, Lineberger Comprehensive Cancer Center, UNC Chapel Hill
  • Dennis Simpson, Lineberger Comprehensive Cancer Center, UNC Chapel Hill

Project Team

  • Kevin Gamiel
  • Clark Jeffries
  • Dave Bowman
  • Xiaojun Guan


Lineberger Comprehensive Cancer Center, UNC-Chapel Hill
Center for Environmental Health and Susceptibility, UNC-Chapel Hill
National Center for Supercomputing Applications


RENCI teams with Carolina Medical Researchers to Develop Better Bioinformatics Tools
IDEA Tools at Dr. Kaufmann’s lab at the UNC Linberger Cancer Center

Melanoma Project Website

Systemic biological models enable assessment of the interactions and complex interplay between the environment and basic biological processes. A biologically faithful model of the human systems of response to DNA damage is of considerable value to environmental and biomedical science. Mitigation of environmental carcinogenesis requires thorough understanding of the complex interactions among DNA repair systems and cell cycle checkpoints to reduce genotoxicity following carcinogen exposure. Cancer chemotherapy, human aging and birth defects are all influenced by DNA damage response. Clinicians and public health managers will design more effective therapies and intervention strategies when the complex organization and structure of the systems of response to DNA damage are established. The great number of public databases filled with genomic and biological function information provides researchers with great opportunities for discovery based science. An automated protocol for interrogating and visualizing interactome models is of broad applicability in the biomedical sciences.

The cyberinfrastructure challenge in interdisciplinary computational biology is to integrate tools and techniques from the physical sciences and engineering with biological and biomedical studies and models. The UNC Program in Melanoma combines medical science expertise from the UNC Lineberger Comprehensive Cancer Center and the Center for Environmental Health and Susceptibility, with expertise in applied mathematics and computational biology to achieve a systems biology level of understanding of an important environmental cancer. To leverage the NCSA high-performance computing infrastructure and domain expertise in computational science, applied mathematics, biochemistry and biology, this project coupled physical science and biological researchers and tools to create state of the art simulations of the human system of response to DNA damage on NCSA clusters.

RENCI, in collaboration with UNC Lineberger Comprehensive Cancer Center, worked to develop tools and software to integrate the large amount of protein-protein interaction data collected in databases (BIND, REACTOME, HPRD, CPATH, etc) and gene expression data generated for this study, and visualization tools to study the systems response to DNA damage based on the integrated genomic data.


Intra-S Checkpoint Interaction Networks

Protein interactons were obtained from the BIOGRID database for human proteins that are part of the intra-S checkpoint and for yeast homologues of the humna proteins. The upper images show the entire networks of protein interactions, the bottom images show the networks of proteins that interact with the replication fork protection complex composed of Timeless, Tipin and Claspin.


Protein Interaction Database

Project team members developed the IDEA database to generate protein interaction network for one species (e.g. human) based on protein interaction data from another species (e.g. yeast).

Interaction Query Tools

To assist in querying the database and to display the resultant interaction networks, graphically, the project team developed a set of IDEA tools.

Note: The current set of IDEA tools are hosted at Dr. Kaufmann’s lab at the UNC Linberger Cancer Center (

Analysis Results

Pathway analysis results for gene lists generated from microarray analyses are available here.


  • Dr. William Kaufmann, Lineberger Comprehensive Cancer Center – email
  • Dr. Xiaojun Guan, Renaissance Computing Institute – email

Acknowledgment for Support

This material is based upon work supported by the National Science Foundation under Grant No. CA SCI 0525308 and CSA SCI 0525308 through the University of Illinois-Urbana/Champaign.


Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.


To aid in the creation of information needed for interaction networks, RENCI developed the Interaction Database for Experimental Analysis (IDEA) for capturing interaction properties.

The database was designed to allow a user to query interactions for a single species, or to build a derived, hypothetical, interaction network for a species (e.g. human) based upon interaction data from other species (e.g. yeast), through homologue mapping.

Input into the database is via standard PSI-MI XML format files representing the interactions. To aid in the retrieval of interaction network information, the project team also developed a set of tools that allow a user to query the database and to display the resultant networks graphically.

Note: The current set of IDEA tools are hosted at Dr. Kaufmann’s lab at the UNC Linberger Cancer Center (