GenProMarkers™, Inc.

Microarray, Bioinformatics and Systems Biology for Discovery of Gene and Protein Biomarkers

Bioinformatics

Bioinformatics is the application of statistics and computer science to the field of moleuclar biology.  Microarray bioinformatics focuses on the application of mathematics, statistics and computer science in the design and operation of DNA microarray experiments and the performance of data analysis. 

Expression Database

To identify differentially expressed genes, we customized computational procedures including data evaluation, filtering, inclusion, normalization, and comparison.  Data evaluation is aimed at revealing data quality including background, signal-to-noise ratio, overall spot pixel intensity, and consistency.  Data filtering is aimed at removing (with uniform criteria described in Methods) all spots that may lead to false positives or negatives.  Data inclusion is aimed at determining whether all spots fulfilling the criteria for data filtering should be used for further analysis.  Data normalization is aimed at removing non-biological variations that can arise in any steps of microarray experiments, including array printing, RNA preparation, labeling, hybridization, or scanning.  Data comparison is to compute averages, standard deviations, ratios, p-values and false discovery rates of the normalized spot background-subtracted mean pixel intensities between desired comparisons, in order to reveal disease-related dysregulated genes. 

Relational Database

Microarray bioinformatics relational database - GenProMarkers Inc.

Overview of database structure. Database includes individual expression files, a relational expression file, a relational gene information file, and category gene information files. Each individual expression file keeps raw microarray data imported from image scanning of a microarray hybridization based on design of array experiments and is linked with the relational expression file via index (a unique numerical ID to each spot). The relational expression file contains information on array design, expression data and comparison, and some gene information.  The relational gene information file contains results of expression comparisons and gene biological information and is linked with the relational expression file via the NCBI Entrez Gene ID.  All or partial information in the relational gene information file can be exported into category gene information files for category analysis and interpretation of results.  A design of microarray experiments, analysis of data, and interpretation of results need information from public databases.  Arrowheads indicate the direction of information flow.  Lines with dots at two ends indicate links.

Individual Database

Microarray bioinformatics Individual Database - GenProMarkers Inc.

Individual database files: (A) Individual expression file, (B) Relational expression file, (C) Relational gene information file, and (D) Category gene information file.

Reproducible microarray findings are guaranteed by experimental design including the technical triplicate (three measurements per gene probe), the experimental triplicate (three microarray experiments per sample), and appropriate biological repeats (the sample size or power calculation).

Microarray Experiments on one pair of test and control samples

Microarray bioinformatics - Microarray Experiments on one pair of test and control 
		samples - GenProMarkers Inc.

Normalization of technical and experimental triplicate microarray data. The quantile normalization algorithms were used to adjust the natural log of the background-subtracted mean pixel intensities of each and every set of 577 genes that were selected from the hMitChip3 triplicate microarray experiments by the criteria described in our publications. In contrast to (C) the prenormalization boxplots, (D) the postnormalized boxplots distribute in the same intervals with the same density center, indicating a successful location adjustment. The postnormalized data were used for further analysis to identify differentially expressed genes with statistical significance (P value < 0.05).

Microarray experiments of human malignant melanomas with technical triplicate, experimental triplicate and biological repeats

Microarray bioinformatics - Microarray experiments of human malignant melanomas with 
		technical triplicate, experimental triplicate and biological repeats - GenProMarkers Inc.

Normalization of Microarray Data.  The quantile normalization algorithms were used to adjust the values of the background-subtracted mean pixel intensities of triplicate measurements per microarray for each and every set of 1037 genes.  Microarray experiments for most samples were conducted in triplicate, excepting for A375, COLO829 and HTB65 in 4 times, and for RPMI7951 and UACC827 in 6 times.  In contrast to the pre-normalization boxplots (top panel), the post-normalized boxplots distribute in the same intervals with the same density center, indicating successful adjustment of data.  The post-normalized data were used for further analysis.  Cell: cell lines; Exp: independent microarray experiments. Log: log2.

Hierarchical Clusters of Genes and Samples

Microarray bioinformatics - Hierarchical Clusters of Genes and Samples - GenProMarkers Inc.

Hierarchical Clusters of 154 differentially expressed genes in UAC903(+6) and UACC903 cells before and after UV treatment.  The “0” indicates that the cells were harvested with the mock treatment without UV irradiation.  Other cells were harvested at 1.5-, 3-, 6- and 12-hour time points after the UV treatment.  The samples were arranged on the top as such to reveal a symmetrical comparison between UACC903(+6) and UACC903 and a timing course for each cell line after the UV treatment.  The expression ratios between UACC903(+6) and UACC903 (+6/903 and 903/+6) were calculated and placed in the middle for the symmetric clustering analysis.  The ‘0/0’ is calculated by dividing an individual value with the mean value at the 0 time point, which provides the control reference (black) for up- (red) and down- (green) regulation or no-change in expression (black) of a gene.  The UV-induced expression changes were calculated by comparisons of a gene expression level at each time point with the mean value at the 0 time point, that is, 1.5/0, 3/0, 6/0 and 12/0.  The triplicate experiments (indicated as 1, 2, 3) were performed for each time point.  The clusters without UV treatment are indicated by (A) 68 genes with expression levels significantly (p<0.05) higher in UACC903(+6) cell line than in UACC903, (B) 47 genes without significant (p>0.05) changes between these two cell line, and (C) 39 genes with expression levels significantly (p<0.05) lower in UACC903(+6) than in UACC903.  The sub-clusters (A-1 through A-4, B-1 through B-3, C-1 and C-2) indicate the UV-induced changes in expression.

Differentially Expressed Genes

Microarray bioinformatics - Differentially Expressed Genes - GenProMarkers Inc.

Heat map of 16 pro-apoptotic, 4 anti-apoptotic, 5 DNA-damage-repair and 5 G2/M-phase control genes in UACC903(+6) and UACC903 cell lines before (0) and at 1.5-, 3-, 6-, and 12-hours after the UV treatment.  The color map illustrates a color change corresponding to a ratio change.  Each ratio was calculated by dividing an expression level of a given gene at the time point to the mathematic mean value of the gene expression levels in both UACC903(+6) and UACC903 cells without UV treatment.  The arrowheads and hyphens indicate overall expression changes of a gene in response to the UV treatment.