Vizxlabs GENESIFTER 2005 User Manual

Download

User’s Guide

GeneSifter Overview

• Login

• Upload Tools

• Pairwise Analysis

• Create Projects

For more information about a feature see the corresponding page in the User’s Guide noted in the blue circle ( ) .

Upload Tools

Upload microarray data files.

Scatter Plot

Interactive scatter plot provides visualization of the entire array data set and identification of individual genes.

Ontology Report

Summarize ontology terms for a gene list and assess the biological significance of the genes within the list.

User Login

Access your secure account from any computer (PC or Mac) with Internet access.

Create New Project

Create user-defined projects with two or more groups. See next page for project analysis options.

Pairwise Analysis

Define two groups and apply normalization, statistical analysis and quality metrics to create lists of differentially expressed genes.

One-Click Gene Summary™

Provides a synopsis of the most current information available for the genes on your array. It includes information from UniGene, LocusLink, Gene Ontology terms and more.

Export Results

Export data and gene annotation to Excel.

GeneSifter Overview

• Project Analysis

• Filtering

• Function Navigation

• Pattern Navigation

•Clustering

Filtering

Apply fold change cutoffs, statistical analysis and quality metrics

to create lists of differentially expressed genes.

Clustering

Identify patterns of gene expression with unsupervised clustering functions.

Ontology Report

Summarize ontology terms for a gene list and assess the biological significance of the genes within the list.

Project Analysis

Project analysis functions allow analysis across all conditions in a project.

Pattern Navigation

Define and identify patterns of gene expression with supervised clustering.

Function Navigation

Rapidly identify and group genes based on function using Gene Ontology terms.

Ontology Report, Cluster

Samples and the One-Click Gene Summary are available for all types of project analysis.

Cluster Samples

Use hierarchical clustering to determine the relationship of samples based on a gene list.

One-Click Gene Summary™

Includes information from UniGene, LocusLink, Gene Ontology terms

and more.

GeneSifter

Introduction and Login

Welcome to GeneSifter, the webbased microarray data management and analysis system, which relies on VizX Labs’ BIOME™ bioinformatics software engine. This document gives an overview of some of the features available in Genesifter. To get started from www.genesifter.net these steps:

1. Select the Login button from the

top right corner.

2. Enter your user name and password in their respective prompts and click on the Login button.

4. A successful login should show a screen with control panel on the left and the most recent announcements concerning Genesifter.

please follow

Genesifter Support and Sales

E-mail: support@genesifter.net Toll-free: 1-877-WEB-GENE Direct: 206-283-4363

Genesifter

Online Help

Genesifter provides page-specific online help.

1. Click on the help icon ( ) to access page-specific help documents. The help icon can be found in the upper right corner of most pages.

2. Clicking on the help icon will open a new browser window which will list the help available for that particular page. Select the document you wish to view.

Uploading Data

Upload Tools

1. In order to upload data, select

Upload Tools from the control panel on the left.

2. GeneSifter offers four tools to load data. The application you use will depend on the format and origins of the data being loaded. QuickLoad Wizard,

Batch Upload, FlexLoad Wizard and Advanced Upload Methods are further described

in the following pages.

Uploading Data

Using QuickLoad Wizard

Use the QuickLoad Wizard to load your data into GeneSifter. Supported platforms for this tool include:

• Affymetrix (native CHP files, or CHP files saved as tab-delimited text)

• Codelink

• Pathways™ 2 & 4

• Spot-On

• Mergen arrays scanned with GenePix®

(all other GenePix files may be loaded using FlexLoad)

Data Files may be archived (zipped) prior to upload.

1. Select Upload Tools from the

control panel on the left.

2. Select Run QuickLoad Wizard from

the Upload Tools page. A new window will guide the user in the upload process.

Uploading Data

Using QuickLoad Wizard (continued)

3. Select the array manufacturer or the image analysis software used and then click the Next button.

4. Select your array from the list of available arrays. If your array is not listed, please

contact scientific support for information on adding the array.After you have selected your array, click Next.

Note for Affymetrix Users: Auto Column

Detection should work for data from MAS 5 and GCOS.

Uploading Data

Using QuickLoad Wizard (continued)

5. Now you will enter information about

the sample (referred to as “Target”) that was hybridized to the array. If you have already entered information about the target, select it from the Select Target pull-down menu, otherwise enter your target using Create New Target.

6. If you create a new target, you will need to select an appropriate “Condition” from the pull-down menu. Otherwise, enter your condition using Create New Condition.

7. Select Next when you have entered all needed information.

8. If your array has two channels, repeat steps 5, 6, and 7 for the second channel.

Uploading Data

Using QuickLoad Wizard (continued)

9. Select Browse and find your data file on your local computer. Select the file and then click the Next button to upload the file to GeneSifter.

10. You will now see a summary of the information you have provided. You can enter a description for the experiment(s) being uploaded. Select Save Data to save the data in your GeneSifter account.

11. When the data is successfully uploaded, you will see Success! as the Last Upload status. Common reasons for failure include: data not in the correct format, or selecting the wrong array at step 4.

12. After saving, you can either load more data by selecting Next or exit the

QuickLoad Wizard by selecting Done.

Uploading Data

Using QuickLoad Wizard

Affymetrix Manual Column Detection

The Upload Wizard will automatically detect the proper data columns for text files generated from MAS 4, MAS 5 and GCOS. If auto detection fails, you can use Manual Column Detection.

1. Affymetrix MAS 5 data format. This is an example of a file from the U34B GeneChip®. Formats may vary due to differences in export from MAS 5. Generally the first column will contain the probeset ID. This identifies each gene on the array. The two columns that are needed by GeneSifter for analysis are the columns containing the signal and detection value for that probe set. In this example, column B, which is labeled Signal, contains the derived signal value and column C, labeled Detection, contains the quality value. There may be additional columns present in a data file and the heading for the signal and quality values may be different than what is presented here. In general the signal column will either be labeled Signal or will have the word Signal at the end of the column name. This column can contain both positive and negative numbers. The quality column will generally be labeled Detection or will have the word Detection at the end of the name. This column will contain the letters A, M and P.

Uploading Data

Using QuickLoad Wizard

Affymetrix Manual Column Detection

(continued)

2. Select Run QuickLoad Wizard as usual (see preceding description). Select Manual from pull-down menu.

3. Enter information about columns in the data file. In the sample file the probeset ID is in the first column (column A), the signal is in column B and the detection call value is in column C, so A would be entered for Probeset Column, B would be entered for Signal Column and C would be entered for Detection Call Column. In the sample file the data begins on the second line of the file so 2 is entered for Data starts on line.

4. Select Next and continue as usual for QuickLoad Wizard. The setting will be saved and used to correctly upload the data.

Uploading Data

Using Batch Upload

Use Batch Upload to load multiple data sets stored in a spreadsheet as a tabdelimited text file. Note: See step 7 for a description of required file format.

1. Select Upload Tools from the control panel on the left.

2. Select Run Batch Upload.

Uploading Data

Using Batch Upload (continued)

3. Enter a name and description for the array you are uploading. The pull-down menu has options for the type of data being loaded including:

Use Affymetrix Probeset IDs – use if the first column of your file contains Affymetrix Probeset IDs instead of GenBank accession numbers.

This File is a GEO Data Set – use if you downloaded data from GEO as a GEO dataset.

Use CodeLink Quality Values –

use if your file has the CodeLink flags G, M, L.

4. Browse your computer to find the file containing your data. Use the Select Array pull-down if you have previously loaded data from this array and you wish to add to that data set.

5. Select Upload Data.

6. Data uploaded. You can either exit Batch Upload by selecting

Done or select Upload More Data.

Uploading Data

Using Batch Upload (continued)

7. The file to be loaded must be a tabdelimited text file (txt). GeneSifter does not accept Excel spreadsheets (xls).

The first column figure) should contain an identifier for that gene. Accepted identifiers are:

 Accession Number  IMAGE Clone ID  Affymetrix Probeset IDs

The second column (Column B) is for an internal identifier. This is left to the discretion of the user and can be left blank.

The next 2 columns contain the intensity (Column C) and quality values (Column D) for that gene in the first experiment to be loaded. Additional experiments are added in the same way (one column for intensity, one for quality).

(Column A in the

The first row A: This cell can be empty. B: This cell can be empty. C: Target name for Experiment #1. D: Condition for Experiment #1. E: Target name for Experiment #2. F: Condition name for Experiment #2,

etc.

must contain this information.

Uploading Data

Using Batch Upload (continued)

8. File format for Batch Upload using

housekeeping genes. The format is the same with one exception: the third column must be labeled HKG (housekeeping genes). The genes that are designated as housekeeping should be marked with an x in this column. The intensities and quality values should follow as stated for Batch Upload without housekeeping genes.

In this example the genes in rows 7 and 9 (TRIM9 and GOLGB1) have been designated as housekeeping genes.

Uploading Data

Using FlexLoad Wizard

Use the FlexLoad Wizard to load data in GeneSifter if the array you are using is not included in the QuickLoad Wizard. Familiarity with the layout of your files is advised before going any further. You will be asked:

• to provide information about the file structure, e.g. what column describes absolute intensity, background intensity, etc.

• how you want the data transformed, e.g. preserve channel intensities or express as a ratio of the two channels.

• how you want the data normalized, e.g. LOWESS

1. Select Upload Tools from the

control panel on the left.

2. Select Run FlexLoad Wizard.

Uploading Data

Using FlexLoad Wizard (continued)

1. The Protocol Title is the name

given to a protocol (i.e. the settings for a specific type of file to be uploaded). You can select a protocol you have already generated or create a new one.

2. If creating a new protocol, replace

“Untitled Protocol” with a Protocol Title.

3. Enter an optional Description for

the protocol.

4. Click on Create New to begin the

creation of a new protocol.

Uploading Data

Using FlexLoad Wizard (continued)

5. If you previously loaded the array into your account, select it from the menu list. Alternatively, if you are creating a new protocol, enter the name of the array in the Create New Array field.

6. Select the number of Channels.

7. Enter the number of files you will be uploading (the maximum allowed at any one time is 30).

8. If you know that the genes are all listed in the same order in every file (experiment) then select Same Order.

If you select Unique IDs, FlexLoad will not assume identical gene order in each file, but instead will utilize a supplied unique identifier. If Unique IDs is selected, every ID for that array must be unique and there cannot be any blank data rows. Ideally, you should use a GenBank Accession Number or an Image Clone ID as the Unique ID to assist in populating the One-Click Gene Summary.

Uploading Data

Using FlexLoad Wizard (continued)

If your data has two channels, from Step 6:

9. Select how you want your data

represented:

Intensities

Data for each channel will be stored separately in GeneSifter.

Ratios

Generates a ratio of the intensities of the red and green channels. GeneSifter only saves the ratio to your account.

Uploading Data

Using FlexLoad Wizard (continued)

10. Provide the column number or letter that contains the gene ID.

11. Identify the type of gene ID by selecting either Auto Detect,

Accession Number, IMAGE Clone ID, or Other. If Accession Number

or IMAGE Clone ID is selected and the data file contains other identifiers or blank rows, errors may occur. In general Auto Detect should be used.

12. Optionally, indicate the column number for gene annotation, if available.

13. Indicate the row number where the intensity data begins. Do not include the column headings.

14. Provide the column numbers that contain the respective data. For example, if your data file contains the cy3 intensity in column 8 and column 10 contains the background, enter 8 for Col 1 and 10 for Col 2 to specify these values for the cy-3 intensities.

Op refers to the operation to perform between the two values. Op may be used to subtract background, or take the ratio of foreground/background for quality control purposes. It is not necessary to enter any value for Op, or Col 2.

Uploading Data

Using FlexLoad Wizard (continued)

This window appears only if you chose Ratios in Step 9.

15. Perform LOWESS Normalization on the data.

16. Select how you want to normalize the data.

17. Method for calculating the ratio of intensities. Per file basis allows you to take into account any dye-

swap experiments you may have.

Uploading Data

Using FlexLoad Wizard (continued)

18. If your targets already exist, you can select them from the pulldown menu. If you need to create new targets, use Advanced Settings.

19. Select Browse to upload the file(s). If you were uploading more files, additional rows would be present. In this screen, two files are being uploaded.

20. For two-color arrays, if Per file basis was selected in step 17,

indicate whether you want the ratios to be cy5/cy3 (5/3) or cy3/cy5 (3/5).

21. A summary of parameters that have been previously selected can be viewed.

22. Select Advanced Settings if you need to enter the target and condition information, or to change the number of files being loaded.

Uploading Data

Using FlexLoad Wizard (continued)

23. Upon selecting Advanced Settings, you can change the

number of files to be loaded.

24. You also have the ability to “Create New Targets/Conditions”. Enter target and condition information for each file to be loaded in the top portion of the screen. Otherwise, select “Use Pre-existing Targets”.

25. You can save the output as a text file formatted for Batch Upload by selecting Save as File or load it directly into GeneSifter by selecting

Upload Files

Uploading Data

Advanced Upload Methods

Robust multi-array average (RMA) is a method for deriving expression measurements from the probe level data contained in an Affymetrix CEL file.

GeneSifter users have the option to perform RMA or GC-RMA during the upload of CEL files. The normalized data is saved in the user’s account for further analysis.

1. Locate the Affymetrix .CEL files you want to load on your local computer.

2. All the files to be uploaded and transformed using RMA need to be compressed into a single ZIP file for upload.

3. Select Upload Tools from the Control Panel.

4. Click on Run Advanced

Upload Methods to begin.

Uploading Data

Advanced Upload Methods

(continued)

5. Select the normalization method. Select the type of Affymetrix array used in your experiments. If your array is not listed, please contact scientific support (support@genesifter.net help loading your array format.

6. Click the Next button to continue.

7. Click Browse to locate the .zip archive containing the CEL files on your local disk. Please note that all the files to be loaded need to be contained within a single .zip file.

8. Choose how you want to create targets and conditions for the loaded files.

9. Click Next to continue.

) for

Uploading Data

Advanced Upload Methods

(continued)

10. Enter information about the data set, including a title and description. If you chose to create groups for your experiment, you will now be prompted to create titles for each group.

11. Click the radio buttons to place experiments into the appropriate group (in the example shown experiments 3 and 4 are assigned to group 2 – FL). You may also change the names for the experiment and target/sample at this point. Click Next to continue.

12. GeneSifter will now complete the uploading and transformation of the data. If the upload is Successful, you can click Done to exit the wizard. Any error messages will also be displayed here. Please contact us if you receive an error message and would like help resolving the problem.

MIAME

MIAME is the Minimum Information About a Microarray Experiment that should be associated with public microarray data. Genesifter can store MIAME, by providing information through a user-friendly interface. The Genesifter V1.5 MIAME interface is designed for extraction and labeling protocols. Subsequent versions will have an interface where more MIAME can be provided.

1. Access the MIAME interface by

clicking on MIAME under Inventories.

2. If MIAME isn’t present, turn it on under

Preferences.

3. Under Preferences > General, turn

MIAME on.

4. A list of all the MIAME Extraction,

Labeling and Sample Treatment protocols created are shown under

MIAME >> Inventories.

5. Create a new MIAME protocol by

selecting the appropriate type in

Create New.

6. Provide a title for the MIAME protocol

(the protocol in this case PCR Protocol is the title).

7. Provide information about the protocol

by selecting values in the pull-down menus. If a suitable value is not in a menu list, it can be created in the other text box. This value will become part of the menu list when the protocol is created.

MIAME

(continued)

8. Many protocols may have

additional information. This can be included in the Full Protocol Description text box (e.g. providing a complete PCR protocol).

9. Select Create to save a new

MIAME protocol.

10. The newly created MIAME object

should appear under

MIAME>Inventories.

11. To edit or delete a MIAME

protocol click on the protocol and select the appropriate button.

MIAME

(continued)

12. To associate MIAME information with a

target click on Targets under

Inventories, choose and click a Target.

13. Click on Edit and select the

appropriate MIAME object.

Note: All Experiments that have this target, will automatically be associated with the newly created MIAME object.

Pairwise Analysis

Set-Up

Pairwise Analysis allows you to set up two groups of data and look for genes that are differentially expressed between the two groups.

1. Select Pairwise under the Analysis heading from the Control Panel.

2. A list of available arrays will show on the screen, along with a description if it was entered upon upload.

3. Select the Analyze icon ( ) to perform a pairwise analysis for a particular array.

4. Upon selecting the Analyze icon, a list of experiments performed using that array is displayed. The experiments are grouped by condition. Select the experiments for Group 1 by checking the corresponding boxes in the Group 1 column.

5. Likewise, select the experiments for Group 2.

6. For an overview of experiment details, select the Document icon ( ).

7. You can now select Analysis Settings parameters to be used in the Pairwise Analysis.

Pairwise Analysis

Set-Up (continued)

8. Select a normalization method from the Normalization pull-down menu. Options are All Mean, All Median or None.

9. Select a method for determining reliability and consistency between replicates of each gene from the Statistics pull-down menu. Options include t-test (two tailed, unpaired Student T), Welch’s t-test, Wilcoxon (non-paramteric) or none.

10. Select Threshold value from the pull-down menu. This number sets the threshold for display of up or down regulation (e.g. setting to 1.5 would select only genes that are differentially expressed by at least a factor of 1.5 in one group compared to the other).

11. You can select a Quality cutoff value for your data. This is optional and the value is calculated differently for different array types.

Pairwise Analysis

Set-Up (continued)

12. You can choose to display Upregulated and/or Down-regulated

genes.

13. The Data Transformation field allows users to specify the following:

• The data should not be log transformed.

• The data should be log transformed for analysis.

• The incoming data has already been log transformed prior to upload.

Note: Transformation is to log

base 2.

14. Perform your analysis by selecting

the Analyze button.

Pairwise Analysis

Advanced Settings

15. If Advanced Pairwise Settings was turned on in the Preferences section, additional parameters are made available for use in analysis.

16. Select an Upper threshold if you would like to set an upper cutoff for the fold differences between Groups 1 and 2.

17. You may decide to include a Correction factor for multiple testing. Options include

Bonferroni (step adjusted p-value), Holm (step down adjusted p-value), Benjamini and Hochberg (false discovery rate), Westfall and Young - maxT or None. If you

select a correction method, you must also be performing a statistical test.

18. Perform your analysis by selecting the Analyze button.

Pairwise Analysis

Results

1. After submitting pairwise comparison analysis parameters, a list of differentially expressed genes is returned. The genes are ordered by ratio, listing those genes which are most differentially expressed at the top of the list.

2. The colored arrow indicates whether expression is higher (red up arrow) or lower (green or blue down arrow; color dependent on Preferences) in Group 2 compared to Group 1.

3. The total number of genes (results found) is indicated next to the Search button. The default gene list display is twenty genes per page; to show more, increase the number in the Show pull-down menu and select Search.

Sort by p-value

If a statistical test was included in the pairwise comparison, the list can be sorted by p-value from the Sort By pull-down menu and then selecting the Search button. Genes will be sorted by ascending p-value. If multiple corrections, have been applied, that p-value is used.

p Cutoff

The user can view results that are less than or equal to the selected p cutoff. A statistical test (t-test, Welch’s t-test or Wilcoxon) must be performed in order to use the p

Cutoff.

Pairwise Analysis

Results (continued)

4. Select the Export Results link to export the results of the pairwise comparison. Depending on your Preferences, data can be exported directly into Excel or saved as a tabdelimited text file.

5. To view the One-Click Gene Summary about any gene in the list, select the gene name.

6. To view more genes within the list, click on the [range] hyperlink.

One-Click Gene Summary™

The One-Click Gene Summary is a powerful feature in GeneSifter that provides a synopsis of the most current information for each gene on an array.

Each gene within a gene list has additional annotation, curated from various data sources.

1. Click on a gene of interest from the results list.

2. A gene summary appears in a new window. Within the upper panel, the mean expression values for each condition (group) are displayed, as well as the raw data for the individual arrays. A link allows you to “Search For Genes With a Similar Pattern”.

One-Click Gene Summary™

(continued)

3. In the lower pane of the One-Click Gene Summary:

Accession No

This ID is from the GenBank database. Clicking on it will display the GenBank record. Specific arrays may have additional IDs, such as Probe Set ID for Affymetrix (links to NetAffx™).

Cluster ID

This ID is from the UniGene database. Clicking on it will display the full UniGene record.

UG Title

The title of the gene cluster in UniGene that contains this gene.

Gene ID

This is the gene Nomenclature ID as curated by the Nomenclature Committee. Clicking on it will display the GeneCard™ record for the gene.

Homologene

Lists genes from the same or other organisms that show high sequence similarity.

Chromosome

Indicates the chromosome on which the gene is present (from UniGene).

One-Click Gene Summary™

(continued)

Cytoband

The cytogenic marker to which the gene belongs (from UniGene).

Seq Count

The number of sequences in the UniGene cluster.

LocusLink

The ID from the LocusLink database. By clicking on it, the LocusLink entry is displayed.

Gene Name

The gene name (taken from LocusLink).

OMIM™

ID number from the OMIM database. Clicking on this link will display the OMIM record for this gene.

RefSeq mRNA

ID from the NCBI Reference Sequence database of comprehensive, nonredundant mRNA transcripts. If you click on (FASTA) the sequence appears in FASTA format in a new window.

RefSeq Prot

The ID is from the NCBI Reference Sequence database of comprehensive, non-redundant protein sequences. If you click on (FASTA) the sequence appears in FASTA format in a new window.

One-Click Gene Summary™

(continued)

Summary

A short paragraph summarizing what is known about the function of the gene (from LocusLink).

Gene Ontologies

Gene Ontology™ terms for the gene (from LocusLink). Each term is hyperlinked to show the GO hierarchy.

KEGG Pathways

A link to the Kyoto Encyclopedia of Genes and Genomes (KEGG). It provides pathway information for gene products.

PubGene

Link to the PubGene providing a tool to explore both literature and sequence co-association networks (from PubGene Tools).

Additional Search Features:

- Perform Sequence Analysis

blastn, blastx and primer design tools

- Search PubMed

Use either the Gene ID(default) or enter another term to search PubMed.

- Search for Homologs

Search within other projects for homologous genes.

™ Network Browser

™ Web database and

BLAST®Analysis

A BLAST search can be performed on any gene sequence within the secure GeneSifter environment.

1. Click on a gene from the results list of either pairwise or project analysis.

2. At the bottom of the One-Click Gene

Summary window, click on Perform Sequence Analysis.

3. A new window will open with the selected sequence pasted in a form ready for BLAST analysis. Select the type of BLAST search to perform (blastn or blastx) from the buttons at the bottom left corner of the window.

4. To determine suitable primers for a selected nucleotide sequence, there is a link to primer3, again with the sequence already loaded into the application.

Note: Use blastn to BLAST the sequence

directly against a nucleotide database. Use blastx to BLAST against a protein database. The latter works by translating your nucleotide sequence in all six reading frames (three for each direction) and blasting each result against the protein database.

3 4

Ontology Report

The Ontology Report is available for any gene list created in GeneSifter. In a gene list from either Pairwise or Project analysis, select the link for Ontology Report in the upper right corner.

1. Gene Ontology (GO) terms are divided into three general categories. The default display will be the Biological Process ontologies. To see the report for either Cellular Component or Molecular Function, select the appropriate link.

2. You can view the data as an Ontology Report or Z-score Report.

3. Ontology summary table. Summarizes GO terms for genes in the list (see number 5-8 for additional details).

4. Pie chart of ontology distribution. Summarizes the frequency of GO terms listed in summary table (see number 9 for details).

Ontology Report

(continued)

5. Genes link displays the genes in the list that correspond to the current ontology.

6. GO link displays the Gene Ontology tree structure for a GO term.

7. Ontology Totals:

List – displays how many genes with the

specific ontology term are in the gene list. For pairwise analysis this list is further divided into upregulated (red arrow) and downregulated (green arrow).

Array – displays how many genes with the

specific ontology were measured in the comparison. Typically, this is the number of genes of that specific ontology on the array being analyzed.

5 6

8. z-score. Indicates whether each GO term occurs more or less frequently than expected. Extreme positive numbers (greater than 2) indicate that the term occurs more frequently than expected, while an extreme negative number (less than -2) indicates that the term occurs less frequently than expected (see number 10 for calculation details).

Ontology Report

(continued)

9. The pie chart of ontology term distribution is a display of the frequency of GO terms in the summary table.

10. z-score calculation:

R = total number of genes meeting

selection criteria

N = total number of genes measured

r = number of genes meeting selection

criteria with the specified GO term

n = total number of genes measured

with the specific GO term

Z-score Report

The Z-score Report is generated from the same data as the Ontology Report. It displays those GO terms that have a z-score greater than 2, or less than -2. In a z-score report generated from a Pairwise analysis, a score is calculated for both up and down-regulated genes. Only one of these scores needs to pass to be included in the cut off.

1. Click on Z-score Report to list the

genes by z-score

2. The default view sorts the terms according to the ontology that is most highly represented within the list. Click on List to sort or re-sort the list in this fashion.

3. Alternatively, to sort the list by z-score in:

a) Pairwise - Click on the arrows.

b) Projects - Click on z-score.

Pathway Report

The Pathway Report is available from any gene list created in GeneSifter by selecting the link for

Reports | KEGG.

Note: The pathway report is currently available only

for human microarrays. The pathway source is from the Kyoto Encyclopaedia of Genes and Genomes (KEGG).

1. All pathways from the gene list are shown.

2. The Genes link displays all genes from the list that are associated with the corresponding pathway.

3. The KEGG link displays the corresponding pathway and highlights the appropriate gene(s) from the gene list.

4. Totals:

List – displays how many genes in the

gene list belong to a pathway. For pairwise analysis this list is further divided into up-regulated (red arrow) and down-regulated (green arrow).

Array – displays how many genes

belonging to a pathway were measured in the comparison. Typically this is the number of genes of a pathway that are present on the array being used.

Pathway Report

(continued)

5. z-score. Indicates whether a pathway

occurs more or less frequently than expected. Extreme positive numbers (greater than 2) indicate that the term occurs more frequently than expected, while an extreme negative number (less than -2) indicates that the term occurs less frequently than expected.

6. z-score calculation:

R = total number of genes meeting

selection criteria

N = total number of genes measured

r = number of genes meeting selection

criteria with the specified pathway

n = total number of genes measured

with the specific pathway

Scatter Plot

After performing a pairwise comparison, the data can be viewed as a Scatter Plot with the log intensities for Group 1 plotted against the log intensities for Group 2. This plot displays the data for all of the genes and color codes the differentially expressed genes. The first time Scatter Plot is selected, there may be a few second delay due to the initial intensive calculation.

1. Select Scatter Plot from the

Pairwise Comparison Results

page. A new window will be displayed.

2. Clicking print plot in the blue control

box will print the scatter plot or save it to a file.

3. The middle toggle line button

allows the user to view/hide the identity line.

4. All the data for that comparison will be displayed on the plot (if multiple experiments were included for a group, the log of the mean intensity for the group is plotted). Red spots represent genes that are more highly expressed in Group 2 and green represents genes that are expressed at a lower level in Group 2. Gray spots represent genes that are not differentially expressed based on the criteria selected for the pairwise comparison. Note that the color scheme of spots can be selected within Preferences.

Scatter Plot

(continued)

5. To identify spots, drag the box over the region of interest and select zoom. The chosen region will be magnified in the box in the upper right frame (see 6). If the box titled hide unchanged is selected in the blue control box prior to zooming, the magnified box will only show genes that passed the threshold and statistical parameters. When zooming regions that are dense with spots, hiding the unchanged genes will generate the zoom box in significantly less time.

6. You can identify spots by mousing over them in the zoom box. The name will be displayed at the top.

7. Clicking on a spot will bring up experimental information in the lower right panel.

8. At the bottom right corner of the main scatter plot window, there is a yellow box to Open static scatter plot. Selecting this will generate a full scatter plot that can be copied, pasted or saved and inserted into reports, notebooks, etc.

Export Results

The results generated from GeneSifter can be viewed in Excel or as a tabdelimited text file. You can export results from either pairwise or project analysis.

1. Within Pairwise Analysis, select

Results: Export.

2. Within Project Analysis, select

Results: Export.

Export Results

(continued)

3. If the Export Files setting in Preferences is set to Excel, the exported data will be displayed using Excel.

4. If the Export Files settings in Preferences is set to Text, the data will be displayed as tab-delimited text.

Create Project from Pairwise Analysis

Any gene list generated in Pairwise Analysis can be saved and further analyzed using additional features in Project Analysis.

1. Select Save as Project to from Pairwise Analysis.

2. Provide a Title and Description for the new (sub)project.

3. Provide optional Notes, for example, how the gene list was generated.

Create New Project

A project is a user-defined set of experiments grouped by experimental Condition or by user-defined categories. Setting up a project allows users to analyze expression across two or more groups.

Create New Project Using Conditions:

1. Select Project from the Create New section of the Control Panel. The default for Create New Project is

Use Conditions to create project.

You will see a list of available arrays.

2. Enter a Project Title (required) and Description (optional).

3. Select an array, or arrays, for use in the Project. When you select a single array, you will see a list of experimental conditions that have been examined on that array. If you select more than one array you will see a list of conditions that are common to all arrays selected. Select Continue after an array or arrays have been selected.

Note: An array appears on this list only if it

has experiments with two or more different conditions.

Create New Project Using Conditions

(continued)

4. You can modify Group Name and enter a Description, if desired.

5. Select a normalization method for each array to be included in the Project. Optionally, select Data Transformation method.

6. Select a group of conditions to include in the project. Click on the condition you want to include in your project. Click the > button to move it to the Selected Conditions box on the right. Once conditions have been moved, select a condition and use the Up and Down buttons to reorder it if desired. Conditions can be removed by using the < button. Click Create Group after conditions have been selected and ordered.

Note: The order of the conditions in the

list will determine how the conditions are displayed when the project is analyzed. The first condition in the list will be treated as the control value. Expression values for other members of the project will be expressed relative to this condition. In the example shown, Follicular Lymphoma (FL) will be the control.

Create New Project Using Conditions

(continued)

7. Select individual experiments to include for each experimental condition. Click the check box to include an experiment. Select Create New Project when all desired experiments have been selected. The values used for analyzing a project will be the mean of all the experiments selected for that condition.

8. You can now add another group, analyze, or create a new project by selecting the appropriate link from the list of choices.

9. The project is summarized and the control is marked with [C].

Create New Project Using New Categories

This feature allows users to create new categories for grouping experiments when creating a project. Users do not have to group experiments by Condition using this

method

1. After selecting Project from the

2. Enter a title for the project

3. Select an array from the Arrays

Create New menu in the Control Panel, select the Create new categories for projects link at the upper right of the page.

(required) and a description (optional).

pull-down menu. Enter the number of new categories you wish to create. Select whether to use the target names for the categories. This option allows you to create a project where each array is analyzed individually, or create a project using both categories and target names.

Create New Project Using New Categories

(continued)

4. Select a normalization method, , data transformation (optional) and enter titles for each of the categories you are creating. Select Continue.

5. All of the experiments associated with the array you have chosen will be displayed. Assign categories using the Category pull-down. You do not have to assign a category to each of the experiments, only those you want to include in the project. Select Continue.

Create New Project Using New Categories

(continued)

6. Select the experiments you want to include by checking the corresponding boxes, or select all experiments by clicking the Select

All Runs link. Select Create Project to finish.

Boxplots

Simple boxplots are commonly used to summarize five aspects of a data set: the minimum, 1 maximum. Boxplots are useful when comparing two or more sets of data. Differences in the median and the spread of the datasets are clearly visible with a boxplot. In addition, the presence of outliers can often be inferred from boxplots.

quartile, median, 3rdquartile and

1. Boxplots are available for “Projects” and

“Sub-Projects” and can be accessed in the Project Details view.

2. In the Project Info area, selecting the

“Boxplot” link will open a new browser window containing the boxplot for the project.

3. The x-axis indicates the “Conditions” (or

“Categories”) in the project. If more than one experiment was included for the condition the graph represents the distribution of the combined dataset.

4. In the Group Info area, boxplots can be

accessed by selecting the “View Boxplot” link for a specific “Condition”. The Boxplot summarizes the combined individual experiment for each of the “Conditions”.

5. The x-axis indicates the individual

experiments included in the project for the selected condition. The boxplot summarizes the data distribution for each of the replicates in the condition.

Boxplots (continued)

Interpreting boxplots

6. The labels on the figure indicate the five attributes of the dataset that are represented in the boxplot. The 1

quartiles indicate the inter-quartile range for the dataset. 50% of the data values lie within this range. The two datasets represented in the figure show very similar ranges (inter-quartile range) and centers (medians).

Note: Genesifter uses “simple” boxplots, i.e.-

the “whiskers” are drawn out to the minimum and maximum value found in the dataset.

and

Maximum data value

quartile for dataset

Median for dataset

quartile for dataset

Minimum data value

Project Analysis

Gene Navigation

Gene navigation allows you to view the expression profile from selected genes in your project. This is the default analysis method for projects. There are three ways that genes may be selected.

1. Search by Name

Enter a gene name or part of a gene name in the text box and search for genes in the selected project.

2. Search by Accession, UniGene, or Probeset ID

Search the current project for a list of Genbank Accession numbers, UniGene cluster IDs, or Affymetrix Probeset IDs.

3. Search by Chromosome

Searches for and displays genes on the selected chromosome.

Project Analysis

Gene Function

Clicking Gene Function in the Project Analysis tool bar, the Search by Gene Ontology and Search by KEGG Pathway options are available.

1. Search by Gene Ontology

Select an ontology from the pulldown list of gene ontology terms. Only those ontologies present on your array are displayed in this list. Note that you can search within Biological Process (default), Cellular Component, or Molecular Function.

2. Search by KEGG Pathway

Select a KEGG pathway from the pull-down list to search for genes that are part of that pathway. The list only contains pathways for genes present on your array.

Both search options allow you to filter

the gene list (Mask or Statistics) and change the gene list display.

Project Analysis

Pattern Navigation

Pattern Navigation allows the user to search for genes that match a user-defined expression profile. For example, this type of analysis could be used to identify genes that are highly expressed at the early stage of a time-course study, but might be minimally expressed at a later stage.

1. Select Pattern Navigation.

Project Analysis

Search by Threshold

Search by threshold allows the user to search the current project for genes that are greater than a specified fold change.

1. Select the desired fold change cutoff by using the pull-down menu next to Threshold. As long as one condition passes the desired cutoff, the gene will be included in the results list. Statistical options are also available.

Project Analysis

Search by Gene Pattern

2. Set a pattern using the pull-down menus

for each condition in a project.

2a. The first variable column allows the user

to set the sign (<, >, =) of the equation for each point of the pattern. Each point refers to the relationship of a condition to the control condition.

2b. The second variable column refers to

the ratio desired for the comparison. It can be set to ratios ranging from 0.3 to

2c. If you want all the conditions set to the

same direction and threshold, you can use the “Set All” option rather than setting each condition individually.

3. Select optional parameters. More

information about specific terms can be found in the page-specific help.

4. Click Search to continue.

Project Analysis

Search by Gene Pattern (continued)

5. To edit the pattern and search again

select the Search Pattern button.

6. Profile displays the user-defined pattern indicating condition and expression ratio.

Project Analysis

Unsupervised Clustering

Cluster analysis can be used to group genes together based on their expression profiles. Genesifter offers PAM (partitioning around medoids) and CLARA (clustering large applications). Both are k-medoids methods, which are variations on the popular k-means algorithm.

1. Select Cluster as the analysis option

for project analysis.

2. Select parameters to be used for PAM analysis and then select the Search button.

Project Analysis

Unsupervised Clustering

3. When the cluster analysis is completed a series of graphs will be displayed. These graphs represent the expression profile for the gene that is the center of that cluster. Individual genes within the cluster have similar expression profiles, which are centered around the profile shown.

An average silhouette width is calculated for the entire dataset and this number can guide the user in selection of cluster number.

4. The number of genes in a given cluster is displayed below the graph for that cluster. Select the graph for a particular cluster to view information about the genes in that cluster.

5. A heatmap figure summarizing the expression profiles for genes in the selected cluster will be displayed.

Cluster Samples

Hierarchical Clustering of Samples

1. Select the Cluster: Samples link. This link is available for any gene list generated by project analysis. Clustering will only be performed for projects which include more than two conditions.

2. A new browser window will open and a dendrogram will be displayed.

Cluster Genes

Hierarchical Clustering of Genes

Hierarchical cluster analysis can be used to group genes together based on their expression profiles.

1. Select the Cluster: Genes link.

This option is available for any gene list generated in ‘Project Analysis’.

2. A new browser window will open

containing the heat map and dendrogram of clustered genes. A heat map view of the whole gene list is displayed at the top of the screen.

Saving a Project Result Set

A filtered gene list returned after performing Project Analysis can be saved and further analyzed using different Project features.

1. Select Results: Save to

create a sub-project from Project Analysis.

2. In the new window, provide a

Title and Description for the new (sub)project.

3. Provide optional Notes, for

example, how the gene list was filtered.

Export

GeneSifter offers the ability to export data in a format compatible with several other packages including: EPClust, Cluster, and GeneSpring. EPClust is a webbased clustering tool that allows the user to submit data over the Internet. Cluster is a desktop application that is freely available. Exporting to these packages is only available within Project Analysis.

From the initial project analysis screen:

1. Select Export from the upper right corner.

Export

(continued)

2. Select the File Format for the tool

of interest.

Optional: Select the link to the

EPClust site, or to download Cluster.

Select ANOVA if you want to export only those genes that pass an ANOVA (p<0.05).

3. If EPClust was chosen, select

whether to normalize your data. Likewise, select Log transform if desired.

4. Click Export.

5. You may save the exported data as a tab-delimited text file, or other file type.

6. Select Save.

Create New Condition

Using the Create New section of the control panel you can create additional conditions, targets, and new projects.

1. To add a new condition begin by

selecting Condition from the control panel.

2. Enter the condition’s title, optionally a description, or any notes and click Create.

3. This will then add the condition to the list of conditions that may be used for uploading data and analysis.

Create New Target

1. To create a new target, select Target from the control panel.

2. Enter the title along with a description (optional), then either select the corresponding condition or Create New. Add any notes (optional), and click Create.

3. Any MIAME information that is associated with a target is also displayed.

4. This will add the target to the list of targets that may be used when uploading data or changing the target in a previously loaded experiment.

Preferences

1. Select Preferences from the Control Panel on the left.

Preference options are subdivided into General, Analysis, User and Account Info and can be selected from the tabbed display.

General Preferences

2. General

 Use Folders- Toggles the use

of folders in Inventories to help organize your data.

 Browser Warning- Warns

user of unsupported browser.

 MIAME- Allows creation and

assignment of MIAME protocols.

 Return Results- Choose

length of the default return for gene lists.

 Language-Select text

language (currently English and Japanese are available).

 Color Scheme- Choose colors

to indicate up and downregulation in displays.

Preferences

(continued)

3. Export

• Include Ontology/ Chromosome SummariesIncludes this information in the exported file.

• Export Files- Choose whether exported results are opened with a text editor or Excel.

4. Uploading

• Show Advanced Methods- Enables the Advanced Upload Methods which allows loading of CEL files, and normalization with RMA or GC-RMA.

• Default Array Type- Allows the user to set the Default

Array Type when using the QuickLoad Wizard.

5. Save desired changes in

Preferences by clicking on the Save button in the lower right corner.

Preferences

Analysis Preferences

4. Advanced Pairwise Settings

4a. When Off is selected, only the basic

parameters for Pairwise Analysis are available.

4b. When On is selected for Advanced

Pairwise Settings, the pairwise analysis screen will include these additional parameters:

• Data Transformation- the ability

to log transform or account for data already log transformed

• Upper- maximum threshold for

expression

• Correction- The Bonferroni,

Holm, BH and maxT statistical corrections for multiple testing

Preferences

Analysis Preferences (continued)

5. Extended Project Data

5a. When Off is selected, only a basic data

summary will be displayed for genes within Project Analysis

5b. When On is selected, a detailed data

summary will be displayed for selected genes within the Project Analysis. An additional table will display all data points used in the calculation. In addition to the information in the basic summary, the detailed summary will include the following:

• Grp Min - The smallest mean

intensity for that condition

• Grp Max - The largest mean

intensity for that condition

• Grp Ave - The mean of all

intensities within the condition

• Grp Med - The median of all

intensities within the condition

Preferences

Analysis Preferences (continued)

6. Row Center Project Data displays data

for a selected gene as the intensity for that gene over the mean intensity for that gene across all groups.

7. Quality Cutoff defines whether all

groups must pass the selected quality cutoff or just one of the entire set must pass.

8. Project Gene Title allows the user to

select whether gene titles or accession numbers are displayed in the project analysis gene lists.

Preferences

Analysis Preferences (continued)

Heatmap Figure

9 ,9a Heatmap Figure graphically

displays gene regulation in gene lists. The colors are modulated by the Color Scheme under General preferences.

• Number Each Row If the option is

selected, each row of the Heatmap figure will be consecutively numbered.

• Open In New Window Upon

clicking “View As Image” in the project results page, the Heatmap figure will be opened in a new browser window.

• Image Type Selects the file format

that the Heatmap image can be saved as. Available formats are PNG and JPEG.

• Title Upon clicking “View As

Image” in the project results page, the resulting Heatmap rows can either be labeled by the gene title or by the accession number.

10. Gene Titles

Sets the display to either accession number or the UniGene title within gene lists. Note that Project Gene Title under Analysis preferences will override this selection.

Preferences

User Preferences

11. User Info

Provides contact information for sending secure email within Genesifter.

12. Change Password

To change password, fill in the fields of old and new password.

13. To save desired changes in

Preferences click on the Save button in the lower right corner.

Preferences

Account Info Preferences

14. Bar graph indicates total account space used.

Appendix

GEO Database

Genesifter can upload GEO formatted files from the Affymetrix platform. The Gene Expression Omnibus (GEO) is a public database of microarray data, hosted by the NCBI (http://www.ncbi.nlm.nih.gov/geo/

1. Query GEO accession

number or the DataSet ID.

2.,3 Alternatively, Browse the

database to find datasets of interest by selecting “DataSets”.

4. Select “Geo platform” in the

Sort by pull-down menu to identify datasets created from the Affymetrix platform and thus capable of uploaded into Genesifter.

)

GEO Database

(continued)

5. Browse through the datasets generated using the Affymetrix platform [Note:GPL91 is the Affymetrix Platform Accesion no in GEO]. Select“go to full dataset record”.

6. To download all the experiments of the dataset, mouse over “download” and select “complete DataSet” and save the file.

GEO Database

(continued)

7. Select Batch Upload option within Genesifter from the Control Panel. Enter an “Array title”, “Description” and select “This file is a GEO Data Set” from “Options”.

8. After uploading the GEO file, Genesifter automatically fills the Target, Conditions and Descriptions for each experiment of the dataset. These field values can be changed from Inventories. Select “Save Data” button to complete the GEO dataset import.

Vizxlabs GENESIFTER 2005 User Manual

Specifications and Main Features

Frequently Asked Questions

User Manual