Eric Wen Su

Comparative analysis and integrative classification of NCI60 cell lines and primary tumors using gene expression profiling data

BackgroundNCI60 cell lines are derived from cancers of 9 tissue origins and have been invaluable in vitro models for cancer research and anti-cancer drug screen. Although extensive studies have been carried out to assess the molecular features of NCI60 cell lines related to cancer and their sensitivities to more than 100,000 chemical compounds, it remains unclear if and how well these cell lines represent or model their tumor tissues of origin. Identification and confirmation of correct origins of NCI60 cell lines are critical to their usage as model systems and to translate in vitro studies into clinical potentials. Here we report a direct comparison between NCI60 cell lines and primary tumors by analyzing global gene expression profiles.ResultsComparative analysis suggested that 51 of 59 cell lines we analyzed represent their presumed tumors of origin. Taking advantage of available clinical information of primary tumor samples used to generate gene expression profiling data, we further classified those cell lines with the correct origins into different subtypes of cancer or different stages in cancer development. For example, 6 of 7 non-small cell lung cancer cell lines were classified as lung adenocarcinomas and all of them were classified into late stages in tumor progression.ConclusionTaken together, we developed and applied a novel approach for systematic comparative analysis and integrative classification of NCI60 cell lines and primary tumors. Our results could provide guidance to the selection of appropriate cell lines for cancer research and pharmaceutical compound screenings. Moreover, this gene expression profile based approach can be generally applied to evaluate experimental model systems such as cell lines and animal models for human diseases.

Evolutionary computation for discovery of composite transcription factor binding sites.

Previous research demonstrated the use of evolutionary computation for the discovery of transcription factor binding sites (TFBS) in promoter regions upstream of coexpressed genes. However, it remained unclear whether or not composite TFBS elements, commonly found in higher organisms where two or more TFBSs form functional complexes, could also be identified by using this approach. Here, we present an important refinement of our previous algorithm and test the identification of composite elements using NFAT/AP-1 as an example. We demonstrate that by using appropriate existing parameters such as window size, novel-scoring methods such as central bonusing and methods of self-adaptation to automatically adjust the variation operators during the evolutionary search, TFBSs of different sizes and complexity can be identified as top solutions. Some of these solutions have known experimental relationships with NFAT/AP-1. We also indicate that even after properly tuning the model parameters, the choice of the appropriate window size has a significant effect on algorithm performance. We believe that this improved algorithm will greatly augment TFBS discovery.

Analysis of pathway activity in primary tumors and NCI60 cell lines using gene expression profiling data.

To determine cancer pathway activities in nine types of primary tumors and NCI60 cell lines, we applied an in silico approach by examining gene signatures reflective of consequent pathway activation using gene expression data. Supervised learning approaches predicted that the Ras pathway is active in ~70% of lung adenocarcinomas but inactive in most squamous cell carcinomas, pulmonary carcinoids, and small cell lung carcinomas. In contrast, the TGF-β, TNF-α, Src, Myc, E2F3, and β-catenin pathways are inactive in lung adenocarcinomas. We predicted an active Ras, Myc, Src, and/or E2F3 pathway in significant percentages of breast cancer, colorectal carcinoma, and gliomas. Our results also suggest that Ras may be the most prevailing oncogenic pathway. Additionally, many NCI60 cell lines exhibited a gene signature indicative of an active Ras, Myc, and/or Src, but not E2F3, β-catenin, TNF-α, or TGF-β pathway. To our knowledge, this is the first comprehensive survey of cancer pathway activities in nine major tumor types and the most widely used NCI60 cell lines. The “gene expression pathway signatures” we have defined could facilitate the understanding of molecular mechanisms in cancer development and provide guidance to the selection of appropriate cell lines for cancer research and pharmaceutical compound screening.

Too much data, but little inter-changeability: a lesson learned from mining public data on tissue specificity of gene expression

BackgroundThe tissue expression pattern of a gene often provides an important clue to its potential role in a biological process. A vast amount of gene expression data have been and are being accumulated in public repository through different technology platforms. However, exploitations of these rich data sources remain limited in part due to issues of technology standardization. Our objective is to test the data comparability between SAGE and microarray technologies, through examining the expression pattern of genes under normal physiological states across variety of tissues.ResultsThere are 42–54% of genes showing significant correlations in tissue expression patterns between SAGE and GeneChip, with 30–40% of genes whose expression patterns are positively correlated and 10–15% of genes whose expression patterns are negatively correlated at a statistically significant level (p = 0.05). Our analysis suggests that the discrepancy on the expression patterns derived from technology platforms is not likely from the heterogeneity of tissues used in these technologies, or other spurious correlations resulting from microarray probe design, abundance of genes, or gene function. The discrepancy can be partially explained by errors in the original assignment of SAGE tags to genes due to the evolution of sequence databases. In addition, sequence analysis has indicated that many SAGE tags and Affymetrix array probe sets are mapped to different splice variants or different sequence regions although they represent the same gene, which also contributes to the observed discrepancies between SAGE and array expression data.ConclusionTo our knowledge, this is the first report attempting to mine gene expression patterns across tissues using public data from different technology platforms. Unlike previous similar studies that only demonstrated the discrepancies between the two gene expression platforms, we carried out in-depth analysis to further investigate the cause for such discrepancies. Our study shows that the exploitation of rich public expression resource requires extensive knowledge about the technologies, and experiment. Informatic methodologies for better interoperability among platforms still remain a gap. One of the areas that can be improved practically is the accurate sequence mapping of SAGE tags and array probes to full-length genes.ReviewersThis article was reviewed by Dr. I. King Jordan, Dr. Joel Bader, and Dr. Arcady Mushegian.

DGEM — A Microarray Gene Expression Database for Primary Human Disease Tissues

Gene expression patterns can reflect gene regulations in human tissues under normal or pathologic conditions. Gene expression profiling data from studies of primary human disease samples are particularly valuable since these studies often span many years in order to collect patient clinical information and achieve a large sample size. Disease-to-Gene Expression Mapper (DGEM) provides a beneficial community resource to access and analyze these data; it currently includes Affymetrix oligonucleotide array datasets for more than 40 human diseases and 1400 samples. The data are normalized to the same scale and stored in a relational database. A statistical-analysis pipeline was implemented to identify genes abnormally expressed in disease tissues or genes whose expressions are associated with clinical parameters such as cancer patient survival. Data-mining results can be queried through a web-based interface at http://dgem.dhcp.iupui.edu/. The query tool enables dynamic generation of graphs and tables that are further linked to major gene and pathway resources that connect the data to relevant biology, including Entrez Gene and Kyoto Encyclopedia of Genes and Genomes (KEGG). In summary, DGEM provides scientists and physicians a valuable tool to study disease mechanisms, to discover potential disease biomarkers for diagnosis and prognosis, and to identify novel gene targets for drug discovery. The source code is freely available for non-profit use, on request to the authors.

Systematic drug repositioning through mining adverse event data in ClinicalTrials.gov

Drug repositioning (i.e., drug repurposing) is the process of discovering new uses for marketed drugs. Historically, such discoveries were serendipitous. However, the rapid growth in electronic clinical data and text mining tools makes it feasible to systematically identify drugs with the potential to be repurposed. Described here is a novel method of drug repositioning by mining ClinicalTrials.gov. The text mining tools I2E (Linguamatics) and PolyAnalyst (Megaputer) were utilized. An I2E query extracts “Serious Adverse Events” (SAE) data from randomized trials in ClinicalTrials.gov. Through a statistical algorithm, a PolyAnalyst workflow ranks the drugs where the treatment arm has fewer predefined SAEs than the control arm, indicating that potentially the drug is reducing the level of SAE. Hypotheses could then be generated for the new use of these drugs based on the predefined SAE that is indicative of disease (for example, cancer).

A Flexible Integration and Visualisation System for Biomarker Discovery

Biological data have accumulated at an unprecedented pace as a result of improvements in molecular technologies. However, the translation of data into information, and subsequently into knowledge, requires the intricate interplay of data access, visualisation and interpretation. Biological data are complex and are organised either hierarchically or non-hierarchically. For non-hierarchically organised data, it is difficult to view relationships among biological facts. In addition, it is difficult to make changes in underlying data storage without affecting the visualisation interface. Here, we demonstrate a platform where non-hierarchically organised data can be visualised through the application of a customised hierarchy incorporating medical subject headings (MeSH) classifications. This platform gives users flexibility in updating and manipulation. It can also facilitate fresh scientific insight by highlighting biological impacts across different hierarchical branches. An example of the integration of biomarker information from the curated Proteome® database (http://www.biobase-International.com/) using MeSH and the StarTree® visualisation tool is presented.

Mining Gene Expression Database for Primary Human Disease Tissues

Studies of gene expression in primary human disease tissue often span several years in order to achieve reasonably large sample sizes and to collect patient clinical information making this data particularly valuable. Due to the lack of a central repository, this data has only been available through disparate and non-publicly accessible sources following publication. We developed disease-to-gene expression mapper (D-GEM) as a publically accessible database and data mining toolbox for microarray data of human primary disease tissue. A statistical pipeline has also been implemented to identify genes over-expressed in disease tissue samples in comparison with normal control samples, or genes whose expression values are associated with clinical parameters such as patient survival rate. One potential application of this data is the identification of pathway specific cancer prognosis markers. By applying a novel, gene signatures for cancer prognosis in the context of known biological pathways in cancer development were identified and confirmed.