Biomedical subjects

Debashis Ghosh

Publications and source records attributed to Debashis Ghosh.

12 recordsLinked to original sources

PubMed2003

Cluster stability scores for microarray data in cancer studies.

BACKGROUND: A potential benefit of profiling of tissue samples using microarrays is the generation of molecular fingerprints that will define subtypes of disease. Hierarchical clustering has been the primary analytical tool used to define disease subtypes from microarray experiments in cancer settings. Assessing cluster reliability poses a major complication in analyzing output from clustering procedures. While most work has focused on estimating the number of clusters in a dataset, the question of stability of individual-level clusters has not been addressed. RESULTS: We address this problem by developing cluster stability scores using subsampling techniques. These scores exploit the redundancy in biologically discriminatory information on the chip. Our approach is generic and can be used with any clustering method. We propose procedures for calculating cluster stability scores for situations involving both known and unknown numbers of clusters. We also develop cluster-size adjusted stability scores. The method is illustrated by application to data three cancer studies; one involving childhood cancers, the second involving B-cell lymphoma, and the final is from a malignant melanoma study. AVAILABILITY: Code implementing the proposed analytic method can be obtained at the second author's website.

Algorithms↗

PubMed2003

Statistical issues and methods for meta-analysis of microarray data: a case study in prostate cancer.

With the proliferation of related microarray studies by independent groups, a natural step in the analysis of these gene expression data is to combine the results across these studies. However, this raises a variety of issues in the analysis of such data. In this article, we discuss the statistical issues of combining data from multiple gene expression studies. This leads to more complications than those in standard meta-analyses, including different experimental platforms, duplicate spots and complex data structures. We illustrate these ideas using data from four prostate cancer profiling studies. In addition, we develop a simple approach for assessing differential expression using the LASSO method. A combination of the results and the pathway databases are then used to generate candidate biological pathways for cancer.

Algorithms↗

PubMed2003

Structure of human estrone sulfatase suggests functional roles of membrane association.

Estrone sulfatase (ES; 562 amino acids), one of the key enzymes responsible for maintaining high levels of estrogens in breast tumor cells, is associated with the membrane of the endoplasmic reticulum (ER). The structure of ES, purified from the microsomal fraction of human placentas, has been determined at 2.60-A resolution by x-ray crystallography. This structure shows a domain consisting of two antiparallel alpha-helices that protrude from the roughly spherical molecule, thereby giving the molecule a "mushroom-like" shape. These highly hydrophobic helices, each about 40 A long, are capable of traversing the membrane, thus presumably anchoring the functional domain on the membrane surface facing the ER lumen. The location of the transmembrane domain is such that the opening to the active site, buried deep in a cavity of the "gill" of the "mushroom," rests near the membrane surface, thereby suggesting a role of the lipid bilayer in catalysis. This simple architecture could be a prototype utilized by the ER membrane in dictating the form and the function of ER-resident enzymes.

Amino Acid Sequence↗

PubMed2003

Impact of diabetes on cognitive function among older Latinos: a population-based cohort study.

BACKGROUND AND OBJECTIVES: Type 2 diabetes, which is highly prevalent in older Mexican Americans, may influence cognitive functioning. We examined the association of diabetes with decline in global cognitive function and memory function over a 2-year period. METHODS: Study subjects were derived from an existing cohort of Latinos aged 60 and over in the SALSA project (n=1,789). Statistical analysis was conducted using logistic regression and a generalized estimating equation (GEE). RESULTS: Logistic regression analysis indicated that baseline diabetes was a significant predictor of major cognitive impairment in Modified Mini Mental State Exam (3MSE) (OR=1.68, 95% CI=1.21, 2.34) and word-list test (OR=1.31, 95% CI=0.99, 1.75). GEE analysis showed that there was no significant difference between diabetic and nondiabetic subjects in change of cognitive scores over 2 years (3MSE, mean=-0.58, 95% CI=-1.48, 0.32; word-list test, mean=-0.10, 95% CI=-0.32, 0.11). CONCLUSIONS: More diabetic complications were associated with major cognitive decline among diabetic subjects. Research on long-term impact of treatment for type 2 diabetes is warranted.

Aged↗

PubMed2003

Impact of antidiabetic medications on physical and cognitive functioning of older Mexican Americans with diabetes mellitus: a population-based cohort study.

PURPOSE: The current study was designed to evaluate the utility of antidiabetic medications in affecting changes in physical and cognitive functioning among older Mexican Americans with diabetes over a 2-year period. METHODS: A longitudinal analysis with repeated measurements between 1999 and 2001 was performed in a cohort of Mexican Americans, 60 or older, in the SALSA Project. Statistical analysis was conducted using a generalized estimating equation. RESULTS: For subjects with diagnosed diabetes </= 5 years (N = 381), there was less decline in physical and cognitive functioning over 2 years among subjects on treatment, compared to those without treatment. For subjects with diagnosed diabetes of 5+ years (N = 337), the effect of antidiabetic medications was more significant in preventing the decline in physical and cognitive functioning (ADL: mean in log scale = -0.10, 95% CI = -0.16, -0.04, 3MS: mean = 6.35, 95% CI = 3.23, 9.48). Combination therapy of antidiabetic agents appeared to be more effective than monotherapy in preventing the decline in physical and cognitive functioning for subjects. CONCLUSIONS: Antidiabetic drugs appear to be useful in alleviating the decline in physical and cognitive functioning among older Mexican Americans with diabetes, especially for those with a longer duration of the disease.

Activities of Daily Living↗

PubMed2003

Diabetes as a predictor of change in functional status among older Mexican Americans: a population-based cohort study.

OBJECTIVE: Epidemiological studies have demonstrated that older Mexican Americans are at high risk for type 2 diabetes and its complications. Type 2 diabetes leads to a more rapid decline in functional status among older Mexican Americans with diabetes. This study was designed to examine the impact of diabetes on change in self-reported functional status over a 2-year period among older Mexican Americans with diabetes. RESEARCH DESIGN AND METHODS: We performed a longitudinal analysis with repeated measurements of functional limitations in a cohort of Mexican Americans aged > or =60 years in the Sacramento Area Latino Study on Aging (SALSA). Diabetes was diagnosed on the basis of self-report of physician diagnosis, medication use, and fasting plasma glucose. Functional status was measured by assessment of activities of daily living (ADL) and instrumental activities of daily living (IADL) at baseline and 1 and 2 years. RESULTS: Of 1,789 SALSA participants, 585 (33%) had diabetes at baseline. Diabetic subjects reported 74% more limitations than nondiabetic subjects in ADL (summary score for number of limitations, 0.99 vs. 0.57; P = 0.002) and 50% more limitations in IADL (summary score for number of limitations, 7.83 vs. 5.25; P < 0.0001). The annual rate of increase in limitations of ADL and IADL was 0.046 and 0.033 (log scale) on each scale among diabetic subjects compared with 0.013 and 0.003 (log scale) among nondiabetic subjects (P < 0.0005). Complications of diabetes were found to increase ADL and IADL limitations among diabetic subjects. Longer duration of diabetes was also associated with an increase in ADL and IADL limitations. CONCLUSIONS: There was lower baseline functional status and a more rapid decline in functional status among older Mexican Americans with diabetes versus those without diabetes.

Activities of Daily Living↗

PubMed2002

The polycomb group protein EZH2 is involved in progression of prostate cancer.

Prostate cancer is a leading cause of cancer-related death in males and is second only to lung cancer. Although effective surgical and radiation treatments exist for clinically localized prostate cancer, metastatic prostate cancer remains essentially incurable. Here we show, through gene expression profiling, that the polycomb group protein enhancer of zeste homolog 2 (EZH2) is overexpressed in hormone-refractory, metastatic prostate cancer. Small interfering RNA (siRNA) duplexes targeted against EZH2 reduce the amounts of EZH2 protein present in prostate cells and also inhibit cell proliferation in vitro. Ectopic expression of EZH2 in prostate cells induces transcriptional repression of a specific cohort of genes. Gene silencing mediated by EZH2 requires the SET domain and is attenuated by inhibiting histone deacetylase activity. Amounts of both EZH2 messenger RNA and EZH2 protein are increased in metastatic prostate cancer; in addition, clinically localized prostate cancers that express higher concentrations of EZH2 show a poorer prognosis. Thus, dysregulated expression of EZH2 may be involved in the progression of prostate cancer, as well as being a marker that distinguishes indolent prostate cancer from those at risk of lethal progression.

Biomarkers, Tumor↗

PubMed2002

Meta-analysis of microarrays: interstudy validation of gene expression profiles reveals pathway dysregulation in prostate cancer.

The increasing availability and maturity of DNA microarray technology has led to an explosion of cancer profiling studies. To extract maximum value from the accumulating mass of publicly available cancer gene expression data, methods are needed to evaluate, integrate, and intervalidate multiple datasets. Here we demonstrate a statistical model for performing meta-analysis of independent microarray datasets. Implementation of this model revealed that four prostate cancer gene expression datasets shared significantly similar results, independent of the method and technology used (i.e., spotted cDNA versus oligonucleotide). This interstudy cross-validation approach generated a cohort of genes that were consistently and significantly dysregulated in prostate cancer. Bioinformatic investigation of these genes revealed a synchronous network of transcriptional regulation in the polyamine and purine biosynthesis pathways. Beyond the specific implications for prostate cancer, this work establishes a much-needed model for the evaluation, cross-validation, and comparison of multiple cancer profiling studies.

Adenosine Monophosphate↗

PubMed2002

alpha-Methylacyl coenzyme A racemase as a tissue biomarker for prostate cancer.

CONTEXT: Molecular profiling of prostate cancer has led to the identification of candidate biomarkers and regulatory genes. Discoveries from these genome-scale approaches may have applicability in the analysis of diagnostic prostate specimens. OBJECTIVES: To determine the expression and clinical utility of alpha-methylacyl coenzyme A racemase (AMACR), a gene identified as being overexpressed in prostate cancer by global profiling strategies. DESIGN: Four gene expression data sets from independent DNA microarray analyses were examined to identify genes expressed in prostate cancer (n = 128 specimens). A lead candidate gene, AMACR, was validated at the transcript level by reverse transcriptase polymerase chain reaction (RT-PCR) and at the protein level by immunoblot and immunohistochemical analysis. AMACR levels were examined using prostate cancer tissue microarrays in 342 samples representing different stages of prostate cancer progression. Protein expression was characterized as negative (score = 1), weak (2), moderate (3), or strong (4). Clinical utility of AMACR was evaluated using 94 prostate needle biopsy specimens. MAIN OUTCOME MEASURES: Messenger RNA transcript and protein levels of AMACR; sensitivity and specificity of AMACR as a tissue biomarker for prostate cancer in needle biopsy specimens. RESULTS: Three of 4 independent DNA microarray analyses (n = 128 specimens) revealed significant overexpression of AMACR in prostate cancer (P<.001). AMACR up-regulation in prostate cancer was confirmed by both RT-PCR and immunoblot analysis. Immunohistochemical analysis demonstrated an increased expression of AMACR in malignant prostate epithelia relative to benign epithelia. Tissue microarrays to assess AMACR expression in specimens consisting of benign prostate (n = 108 samples), atrophic prostate (n = 26), prostatic intraepithelial neoplasia (n = 75), localized prostate cancer (n = 116), and metastatic prostate cancer (n = 17) demonstrated mean AMACR protein staining intensity of 1.31 (95% confidence interval, 1.23-1.40), 2.33 (95% CI, 2.13-2.52), 2.67 (95% CI, 2.52-2.81), 3.20 (95% CI, 3.10-3.28), and 2.50 (95% CI, 2.20-2.80), respectively (P<.001). Pairwise comparisons demonstrated significant differences in staining intensity between clinically localized prostate cancer compared with benign prostate tissue, with mean expression scores of 3.2 and 1.3, respectively (mean difference, 1.9; 95% CI, 1.7-2.1; P<.001). Using moderate or strong staining intensity as positive (score = 3 or 4), evaluation of AMACR protein expression in 94 prostate needle biopsy specimens demonstrated 97% sensitivity and 100% specificity for detecting prostate cancer. CONCLUSIONS: AMACR was shown to be overexpressed in prostate cancer using independent experimental methods and prostate cancer specimens. AMACR may be useful in the interpretation of prostate needle biopsy specimens that are diagnostically challenging.

Biomarkers, Tumor↗

PubMed2002

Resampling methods for variance estimation of singular value decomposition analyses from microarray experiments.

Microarray experiments offer the ability to generate gene expression measurements for thousands of genes simultaneously. Work has begun recently on attempting to reconstruct genetic networks based on analyses of microarray experiments in time-course studies. An important tool in these analyses has been the singular value decomposition method. However, little work has been done on assessing the variability associated with singular value decomposition analyses. In this report, we discuss use of the bootstrap as a method of obtaining standard errors for singular value decomposition analyses. We consider use of this method both when there are replicates and when no replicates exist. The proposed methods are illustrated with an application to two datasets: one involving a human foreskin study, the other involving yeast.

Algorithms↗

PubMed2002

Mixture modelling of gene expression data from microarray experiments.

MOTIVATION: Hierarchical clustering is one of the major analytical tools for gene expression data from microarray experiments. A major problem in the interpretation of the output from these procedures is assessing the reliability of the clustering results. We address this issue by developing a mixture model-based approach for the analysis of microarray data. Within this framework, we present novel algorithms for clustering genes and samples. One of the byproducts of our method is a probabilistic measure for the number of true clusters in the data. RESULTS: The proposed methods are illustrated by application to microarray datasets from two cancer studies; one in which malignant melanoma is profiled (Bittner et al., Nature, 406, 536-540, 2000), and the other in which prostate cancer is profiled (Dhanasekaran et al., 2001, submitted).

Algorithms↗

PubMed2002

Singular value decomposition regression models for classification of tumors from microarray experiments.

An important problem in the analysis of microarray data is correlating the high-dimensional measurements with clinical phenotypes. In this paper, we develop predictive models for associating gene expression data from microarray experiments with such outcomes. They are based on the singular value decomposition. We propose new algorithms for performing gene selection and gene clustering based on these predictive models. The estimation procedure using the regression models occurs in two stages. First, the gene expression measurements are transformed using the singular value decomposition. The regression parameters in the model linking the principal components with the clinical responses are then estimated using maximum likelihood. We demonstrate the application of the methodology to data from a breast cancer study.

Breast Neoplasms↗