Colorado Neuroinformatics Tools
URL: http://phenogen.ucdenver.edu

• Upload raw microarray data using the INIA copy of MIAMExpress

MIAMExpress is a web-based tool for gene expression data, that can be used for data submissions to ArrayExpress (see section on “Uploading an Experiment in MIAMExpress” in the INIA manual).  In addition to all of the Affymetrix arrays (for mouse, rat, and human), users can upload CodeLink mouse (UniSet1 and whole genome) and rat (whole genome) arrays manufactured by GE Healthcare.

• Upload gene lists

Users can a upload list of genes (list may include any of the gene identifiers such as official gene symbols, RefSeq ID, GenBank ID, and/or probe IDs) and perform further analysis using any of the tools listed below (see section on “Manually Entering a gene list” in the INIA manual).

• Share data with other INIA investigators

Uploaded data (which is always under the control of the INIA investigator) can be made available to other INIA investigators upon request and consent of the appropriate Principal Investigator (see section on “Security” in the INIA manual).

• Search literature

This tool allows users to perform a literature search (PubMed search) for various genes as well as to search for the co-references for each gene in the list (see section on “Literature Search” in the INIA manual).

• Gene annotations using multiple databases

Users can obtain annotations (basic and advanced) for a list of probe IDs from multiple public databases such as GenBank, Ensembl, MGI etc., and in addition also obtain availability of genetically modified animals, for any of the genes in the list (see sections on “Annotating Gene List” and “Basic Annotation Example” in the INIA manual).   Annotation tables (Basic and Advanced Annotations) also contain expression QTL for the genes in the list (see section on “Expression QTL” in the INIA manual).  The purpose of expression QTLs (eQTLs) is to determine the location in the genome that controls the transcription level of a gene. eQTLs are calculated using traditional QTL techniques where the quantitative trait of interest is the expression level of a gene as measured by microarray analysis.  These mouse eQTLs reported in the Basic Annotation page are derived from whole-brain gene expression data for a panel of 30 BXD RI strains (5-6 male mice of 10-12 weeks of age were used per RI strain).  Probe set intensity values were normalized and summarized using RMA. Mean expression levels within strains were used as phenotypic values in a QTL analysis implemented in QTL Reaper, which is written in C and compiled as a Python module.  A weighted marker regression analysis was used within QTL Reaper to calculate LRS scores for each marker (for further details see section on “eQTL” in the INIA manual).

• Utilizing expression- and phenotypic-QTL data

At present users can obtain information about any behavioral QTLs from MGI (a link to MGI is provided on the INIA website).  Users will also be able to obtain information about co-localization of expression- and phenotypic-QTLs using eQTL Explorer.

• Oligonucleotide microarray quality control analysis

At the Colorado INIA website “quality control checks” are carried out in three different steps.  First “array compatibility” (i.e., whether the arrays selected by the users for data analysis are similar in terms of platform and versions within the given platform) is checked for all the arrays selected by the user.  At the second step MIAMExpress information for these arrays is compared and discrepancies, in any, are listed in the “Quality Control Report” for the users.  In the third step all the Affymetrix arrays are taken through “within array” (five different parameters such as average background, hybridization controls, integrity of transcripts for housekeeping genes, percent present calls and scaling factor) and “between array”  (a model based quality control assessment is carried out for all of the user selected arrays.  This involves assessment of three different parameters - pseudocolor images of all the arrays, relative log expressions and normalized unscaled standard error).   A different sets of Quality control checks (such as relative log expression, coefficient of variation, and the quality control flags used within the software that analyzes CodeLink array images) are used for CodeLink arrays.  A report is generated delineating results of all of the quality control steps (see section on “Quality Control Checks Overview” in the INIA manual).

• Oligonucleotide microarray data normalization

A number of different options, based on the microarray platforms are available to normalize data uploaded using MIAMExpress.  For Affymetrix arrays users can normalize the data using any of the following options –MAS v5, RMA, GC-RMA, dChip, and VSN.  For the CodeLink arrays users can normalize the data using any of the following options – Loess, VSN and LIMMA (see section on “Data Grouping and Normalization” in the INIA manual).

• Data filtering (noise filtering)

A typical microarray consists of thousands of probe sets (10,000 – 45,000). As a result, the introduction of meaningless noise in the statistical analysis of data is inevitable. Removing this noise increases the chances of finding differentially expressed genes. For Affymetrix arrays, users can use “Control Gene filter” and “Absolute Call filter”.  For CodeLink arrays, users can use “Control Gene filter”, “CodeLink Call filter”, “Median filter” and “Coefficient Variation filter” (see section on “Data Analysis” in the INIA manual).

• Statistical analyses, including most common statistical tests and permutations

Users can select either parametric and non-parametric tests and a variety of multiple testing correction options such as Bonferroni, Holm, Hochberg, SidakSS, SidakSD, FDR (Benjamini and Hochberg OR Benjamini and Yekutieli) and Permutation (minP and maxT) are available to carry out statistical analysis of microarray data (see section on “Statistical Analysis Overview” in the INIA manual).

• Promoter analysis (analysis of transcription factor binding sites)

A customized version of “oPOSSUM” is available: this tool allows determination of the over/under-representation of transcription factor binding sites (TFBS) within a set of (co-regulated) genes as compared with a pre-compiled background set (see section on “Promoter Analysis Overview” in the INIA manual).

• Upstream sequence extraction

This tool can be used to extract the upstream DNA sequences for a list of genes.  These sequences can be used carry out analysis of transcription factor binding sites, in a list of co-regulated genes, using any other available tools the users may have available (see section on “Upstream Sequence Extraction Overview” in the INIA manual).