Agilent | Version Notes – GeneSpring GX 10.0.2

Literature Library

Technical Support

Education & Events

GeneSpring GX 10.0.2 - Released 13-Mar-2009

Last Update: 13-Mar-2009

New Features

Automatic notification upon start up of GeneSpring application when Pathway Interaction Database updates are available.
Automatic detection of whether data in Affymetrix CHP files are linear or logged transformed. Each state will then be handled appropriately.

Bug Fixes

In rare cases PLIER and IterPLIER do not converge even after default GeneSpring setting of 3000 iterations. In such cases, an error message is thrown and experiment creation fails. A fix has been provided where the probe sets which do not reach convergence will be assigned a pre-variance stabilized signal value of 0. Additionally, the experiment will have 2 new entity lists created, one having probe sets which converged and the other which has probe sets which did not converge.
Multivariate splicing ANOVA crashed when it encountered transcripts with large number of exons in experiments with an unbalanced designed. This has been fixed.
While creating an experiment using data from Mouse Exon 1.0 ST array, an unknown error occurs upon preprocessing when transcript level = Full is chosen. This has been fixed.
Creation of Biological Genome from NCBI failed because NCBI changed the directory of the ALL Data.gene_info.gz file that GeneSpring was pointing to. GeneSpring is now pointing to the new directory and will be able to retrieve the ALL Data.gene_info.gz file that is necessary for creation of the biological genome.
CHIP_TYPE_NOT found error statement is presented when user imports Agilent miRNA GeneView file. Only Full text files are supported for Agilent miRNA data. The error statement has been changed to more clearly convey this message.
Probeset ID Hyperlinks to NetAffx were broken for Affymetrix Gene ST and Exon ST arrays due to a change in URL. This has been fixed.
Removed -Xss parameter to GeneSpring on Mac OS. This was leading to some sporadic crashes.
In rare cases, in the PCA plot certain samples were being colored wrongly due to incorrect ordering of samples. This has been fixed.
GeneSpring Workgroup client does not start up when it fails to connect to the Pathway Interaction Server. This usually happens when the Pathway Interaction Server is installed on a separate machine from the Workgroup Server. This has been fixed.
If Pathways present within GeneSpring before version 10.0.0 had not been opened even once in 10.0.0 and the user upgraded to GeneSpring GX 10.0.1, then while opening the pathway an error would occur. This has been fixed.
K-means cluster view launched on average interpretation having two or more parameters was showing the condition grouping incorrectly in the view. This has been fixed.

Known Issues

For RTPCR experiments, extraction of Annotations as described in section 14.1.1 in the GeneSpring manual works only if the header of the file has SDS 2.2RQ Results 1 as the software version.
For RTPCR experiments, in some cases the data files having the following supported formats RQ1.2, RQ 2.1 and RQ2.3 might give the error File Format not supported. In such cases the error can be resolved by deleting all the header information from the files until the start of the data header row. After deleting all the headers, the first line in the file should have the columns PlateID, Sample, Detector, Ct Avg etc.

GeneSpring GX 10.0.1 - Released 16-Dec-2008

Last Update: 16-Dec-2008

New Features

If the MySQL server containing the Pathway Interaction database does not start, the user will be informed and will be allowed to perform other GeneSpring functionalities apart from Pathway Analysis.
GPR technology creation is more flexible; it is compatible with different versions of the gpr file containing at least the following columns - ID, F635 Median - B635, F532 Median - B532, and Flags.

Bug Fixes

In the Scatter Plot, the data points at the edge of the axes were not getting plotted. This has been fixed.
Memory optimizations while running IterPLIER for alternative splicing analysis have been put in.
When certain specific entities are present in a list which is used as an input for splicing visualization, incorrect probeset annotations were being shown. This has been fixed.
Memory optimizations while importing large BioPAX files have been put in.
Advanced Search using search condition “starts with” was returning erroneous results. This has been fixed.
In Illumina experiment creation, the order of the probesets in the technology and the experiment created were different; subsequent operations on the experiment like Import Entity Lists from file would have resulted in wrong annotations being imported. This has been fixed.
The Circo algorithm has been removed from the PA Analysis wizard. It is not supported.
Robustness checks while doing GeneSpring GX 7.3 - GeneSpring GX 10 migration have been put in.
For Generic Single Color experiments created from a file with multiple samples, the Filter on Data Files wizard was showing only one column collapsing all the column values into this column. This has been fixed.
Identification of miRNA gene targets was resulting in fewer targets than those that exist in the TargetScan database. This has been fixed.
For certain transcript clusters, splicing visualization incorrectly plotted negative values of the probeset intensity values. This was because of an extra median shift that was being performed on the probeset signals. This has been fixed.

GeneSpring GX 10.0 - Released 15-September-2008

Last Update: 15-September-2008

New Features

Analyses

GeneSpring GX 10.0 has built-in support for Pathway Analysis. Natural language processing (NLP) derived databases of biological and chemical entities and relationships between these entities are downloadable for the following organisms. Human, Mouse, Rat, Arabidopsis, Drosophila, C. Elegans, E. coli and Yeast. Most relationships between entities in these databases are supported by viewable references. Algorithms to query these databases for entities and relationships and the ability to build relevant biological networks from genes of interest has been added. An NLP algorithm has been integrated to extract relationships from PubMed abstracts or documents in PDF, HTML, DOC and TXT formats. A shared server version of the databases (called the Pathway Interactions Server) is also available for Workgroup customers and provides centralized sharing of entities, relationships, pathways and fast database queries. Users can also build their own organism database by creating an empty organism database and then importing BioPAX pathways into that database. A PathwayArchitect import function allows pathways exported as .bin files from PathwayArchitect to be imported into GeneSpring. A collection of about 100 curated disease and signalling pathways from PathwayArchitect have been provided in the samples subfolder; these can be imported into GeneSpring via a PathwayArchitect import function. Finally, a new Pathway Experiment type has been added for easy import and subsequent pathway analysis of a list of genes/small molecules.
Support for analysis of miRNA data using Agilent Full Text Feature Extraction files has been added. TargetScan information has also been added to allow identification of miRNA gene targets.
Support for alternative splicing analysis using Affymetrix ExonChip CEL/CHP files has been added. Probeset and transcript filtering tools have been added to identify entities to be considered for further analysis. A multivariate splicing ANOVA has been added for detection of alternative splicing events. A probeset level view of splicing indices has been added for viewing splicing events.
Support for analysis of Applied Biosystems 7900 real-time PCR SDS and RQ manager text files has been added. To know more about the exact versions please refer to the user manual.
Gene Set Analysis (GSA) algorithm has been implemented.
Gene Set Enrichment Analysis (GSEA) algorithm has been fine-tuned with improvements in algorithms for ranking genes and many speed optimizations.
GenePix Results (GPR) files output by the following GenePix Pro versions are supported - 3.0.6.x, 4.0.1.x, 4.1.1.x and 5.0.1.x.
The Filter on Volcano Plot tool is now implemented in the Analysis section of the Workflow panel.
Find Significant Pathways step has been added to each of the Guided Workflows provided in GeneSpring GX.
An Ingenuity Pathway Analysis (IPA) plugin containing three IPA scripts, Create Pathway in IPA, Perform Data Analysis on Experiment, and Perform Data Analysis on Entity List, has been implemented.
The PCA-based Clustering algorithm has been removed from the clustering algorithm options. PCA analysis though continues to be there as a separate link in the Analysis section of the workflow.

Data Visualization

Fold-change lines have been added to the scatter plot and entities outside or within the fold-change lines can be saved to an Entity List.
MvA Plot is now available for both one-color and two-color data.
The Plot Entity List Associated Values tool has been added. This will allow you to select up to two Entity Lists and plot each entity according to its expression values and/or list associated values saved in the Entity Lists.
Entity Lists from different experiments can now be projected to the same Venn Diagram.
In views such as Profile Plot, Scatter Plot, Heatmap and others, mouse-over on an entity now displays annotations for that entity.
A 3-D scatter plot for PCA on Conditions in QC on Samples and PCA tools has been added.
Improvements to hierarchical tree display have been made. The hierarchical tree now displays a compact birds eye view of the tree on the left, and the complete tree on the right. Selection and zoom is supported by clicks on the tree nodes or by drawing a rectangle on the region of interest. Subtrees and classifications can be created from the tree on the right.

Data Import , Export, and Organization

Many improvements to the GX 7 to GX 10 data migration tool have been made. We have automated many steps that were required to migrate GX 7 data into GX 9 into two simple steps. Select the GX 7 genome(s) you would like to migrate. Upon completion of migration, a GX 10 project will be automatically created for each genome migrated containing all experiments, gene lists, interpretations, parameters, parameter values, samples, trees, and classifications in the genome(s). To view the migrated GX 7 genome and its associated data objects, open the Project name that matches the name of the genome in GX 7. Upon opening the Project, you should see all the migrated experiment(s), along with their associated objects previously mentioned. Migrated GX 7 experiments can also be searched for (Search > Experiments) and added to any other GX 10 Projects. If an experiment was assigned to a project in GX 7, the experiment will be migrated into GX 10 in a GX 10 project that matches the name of the project in GX 7.
Raw intensity values are now shown in Entity Inspector and can be exported in Export Entity List, plotted in Plot Entity List Associated Values, and filtered on in Filter Probesets by Expression. Raw intensity values in GX 10 are defined as the values imported into GeneSpring GX from the data files after applying the user defined threshold and performing summarization. Experiments created in GX 9 will be automatically upgraded to have these raw value representations whenever they are opened (therefore, opening a GX 9 experiment for the first time will take longer than usual)
Summarization of values in all Agilent and Generic Experiments is now performed on the log scale (after applying the user threshold) rather than the linear scale.
A My Favorites folder has been added to each experiment. Entity Lists, trees, classifications, and pathways can be copied from the experiment and pasted into the My Favorites folder. Data objects in My Favorites folder can be organized in any manner. Sub-folders can also be added to the My Favorites folder.
Data that have already been log-base 2 transformed can now be imported into GeneSpring GX 10 to create generic one or two-color experiments, as the option to skip log-base 2 transformation has been added.
Copy to Clipboard capability has been added to most of the data tables in the right-click menu.
The Advanced Search feature has been speeded up substantially. This improves efficiency at several places, in particular, in GSEA where searching for gene sets took substantial time earlier.
Generic Single and Two Color import now supports more extensive support of missing values. Correlation coefficients between arrays are now computed only on genes in which both arrays have non-missing values.
Import of files for Generic single and two color experiments now merges files by identifier rather than merging files by order of occurrence. Multiple occurrence of an identifier are considered in order. However, it is necessary that the first file specified be a superset file.

Technology Annotations

More flexibility in GO ID formats has been added in GeneSpring GX 10. Please see the GeneSpring GX 10 User Manual for details on supported formats.
Technology annotations can be updated directly from NCBI through the Create Biological Genome and Update from Biological Genome functions.
Technology for both catalogue and custom Agilent arrays can be created directly from eArray. Technology annotations can also be updated from eArray.
Translation of Entity List into an experiment of a different technology now allows you to import entity list associated values and raw or normalized signal intensity values from the source experiment into the target experiment. These values can be viewed in the Entity List Inspector of the Entity List created in the target experiment. In addition, these values will be available for use in the Plot Entity List Associated Values tool.
A homology table is now shown when entity list is translated into an experiment of a different technology.
Translation of Entity Lists across organisms using Homologene mapping is now supported for all organisms for which Homologene mappings exist.

Normalization and Baseline Transformation

Normalize to Control Genes has been added as a normalization option for one-color data. Control genes can either be searched for in the technology, or imported from a file.
Scaling has been added as a normalization option for one-color data.
Ability to specify multiple control groups to use for baseline transformation has been added.

Quality Control on Probesets

Filter Probesets on Data File has been added to filter probesets on values in any column of your imported data files.
Filter Probesets on Error has been added to filter probesets based on their calculated standard deviation or coefficient of variation across the replicate samples of the selected interpretation.
Filter Probesets on Expression has been modified to filter probesets based on their normalized or raw intensity values, in addition to the percentile ranking of their values.

Miscellaneous Changes/Additions

Exclude certain flags in Interpretation option has been added so you can choose to only consider measurements marked with specific flags when calculating the normalized intensity values for a probe across replicate samples.
Limit on search results has been removed and all results from a search will be displayed. By default, the number of results displayed per page is 100. This can be changed within the search window.
Show name of the Agilent microarrays- In GeneSpring GX 9, only the AMADID were shown for Agilent microarrays in the Update Technology window. We are now showing the catalogue name of the arrays.

Fixed Bugs

Project containing experiment created from custom Affy CDF when exported as project zip does not export the technology also. One can now specify which technologies to export along with the project zip. Fixed
Import project zip did not work earlier for zip file sizes larger than 2GB because of limitations in the utility used for unzipping. Larger project zips can now be imported via the following mechanism. Use any external utility to unzip and then point GeneSpring to the resulting unzipped folder. Fixed
Coloring of column names in hierarchical condition clustering is not proper in the cases where column name is different from sample name. Fixed

Known Issues

The following issues have been found in 10.0 and will be addressed in a subsequent revision.

Some hierarchical trees built in GX 9 (in particular, those with high depth of nesting) may not open in GX 10. The reason for this is that GX 10 has a different (lower) setting for a Java stack size parameter as compared to GX 9. This new setting is necessitated by the addition of a MySQL database for Pathway Analysis. Opening such trees will result in a message asking users to add a –Xss2m term to the java.options line in installationFolder/bin/packages/properties.txt. Make this setting and then restart and try opening the tree. This may need to revised to 20m for some particularly deep trees. Note that any trees you build in GX 10 do not have this issue and therefore we recommend that you rebuild any problematic trees in GX 10.
Setting the –Xss parameter to 2m or above is known to be associated with sporadic but rare crashes, usually with a StackOverflowError showing in the installationFolder/stderr.log file.
CEL (and CHP) files cannot be in a directory (or path) that contains non-Latin characters (such as Kanji) due to limitations of the Affymetrix File reading API. Ensure that any files that need to be read by GeneSpring are in folders that do not contain any non-latin characters.
Pathway Views are the only editable views which are not auto-saved. Users will need to remember to save these views periodically while editing.
Scatter plots lose axis markings when displaying very small values, in the range 10E-5 or smaller.
In Pathway Experiments, lassoing between pathways views, lists and data views happen purely on the basis of Entrez Ids; therefore, selecting a small molecule on the pathway view will not select that molecule on a scatter plot. Further, creating a new list by selecting on a pathway view will not create a list with all selected entities; only those which have Entrez ids and which are a subset of the starting entity list will find their way into the newly created list.
Pathway Databases require a MySQL server whose speed depends on the amount of memory available. In 64bit 4GB RAM machine, MySQL settings need to be tweaked to provide much faster performance than what is achievable on a 2GB RAM machine.
NLP running on pdfs uses a pdf to text convertor which can handle many but not all kinds of pdf. The same holds for .doc to text conversion.
Shortest connect queries are not guaranteed to connect all the input entities by paths as the algorithm stops when the size of the network it is dealing with becomes very large.
Opening a GX 9 experiment for the first time in GX 10 will take a little longer than usual because of the updation process for raw values. For two color experiments, this will also require an increase in memory needed for storage.
Tiger Mac machines are not officially supported yet because of of a sporadic CreateNewSharedObject error caused by a Mac OS bug; this error typically comes when running Affymetrix summarization algorithms, clustering or layouts on the pathway view.
Hierarchical trees sometime gray out when there are several of these open and another window is moved over them; usually maximizing/minimizing the application will fix the problem.
The score property of a relationship is computed based on only one rather than all of the associated references. This affects only relationships with multiple references which are a small fraction of all the references.

GeneSpring GX 9.0.5 - Released 15-April-2008

Last Update: 15-April-2008

New Features

Workgroup client mode allows for storage of data and results on a centralized server and execution of certain tasks on a compute server cluster.
Entity lists can be imported from a tab delimited, CSV or XLS file.
Library annotations can be updated from a tab delimited, CSV or XLS file.
On startup, GeneSpring GX will check for updates to the product and technologies and will inform the user when there are updates which can be applied immediately.
Technologies can now be deleted.
Samples not used in any experiments can now be deleted.
General speed and memory optimizations for analyzing large experiments.

Fixed Bugs

GO browser does no longer rely solely on EntrezGene ID’s, but can use other identifiers to perform GO analysis.
Deleted pathways from database still used for Find Similar Pathways analysis. Fixed.
Tiff doesn’t show up in the drop down in the file chooser on export as image. Fixed.
Inconsistencies with the commas and periods on European systems that use comma’s for decimal points. Fixed.
The startNetworkServer.sh packaged with the application doesnt work on MacOS X. Fixed.
Excel file based samples cannot be re-used for creating an experiment. Fixed.
GO ids that did not include the leading zero’s could not be parsed out correctly. Fixed.
GSEA - every run of GSEA with same set of inputs can result in potentially different results. The random seed inside the algorithm is not fixed. Fixed.
GSEA - All genesets are converted to Gene Symbol before execution. If it is not possible to convert one of them, they are used as is. However their mark is different and only genesets of one mark are displayed in the results table. This happens in advanced search option only. Fixed.
GSEA - only one of the two result tables "Gene Sets satisfying minimum genes" and "Gene Sets failing minimum genes" show. Fixed.
GSEA - Clicking cancel while "Choosing Gene Sets with Minimum Matches" progress message gives two error dialogs. Fixed.
If you have two objects of the same name in a multi-inspector, on switching between the two, the inspector page for each does not change. Fixed.
A condition tree built on average columns does not show the gene id column as labels. Fixed.
Search - set of returned results from the search is filtered out by the set of results that are not deleted. So reducing the max search limit will sometimes result in confusing results. Fixed.
Condition tree - If you right click and open and if samples are removed, then it shows "Cannot launch" error followed by "0/-1" error. Double click open works as expected. Fixed.
GSEA - Clicking cancel while "Choosing Gene Sets with Minimum Matches" progress message gives two error dialogs in stead of one. Fixed.
GSEA - If you execute once, and then go back on the wizard and execute with a different set of inputs and cancel midways, old results are displayed. Fixed.
GSEA - use advanced search option -> choose lists -> run GSEA -> go back to choose lists page -> list empty.Fixed.

Known Issues

The following issues have been found in 9.0.5 and will be addressed in the next revision.

CEL (and CHP) files cannot be in a directory (or path) that contains non-latin characters (such as Kanji) due to limitations of the Affymetrix File reading API. Ensure that any files that need to be read by GeneSpring are in folders that do not contain any non-latin characters.
Tool crashes, when Hierarchical clustering algorithm runs with one sample, which is created by preprocessing with 'Baseline to median of all sample', in Mac. It works fine in Windows and Linux.
Bug in generic 1-color/2-color wizard - No proper error message if you choose wrong options in selecting rows. For example, choose between marks gives blank error. eg2: if the column header is not selected, gives error "No categorical or continuous columns "
On correlation plot in QC on samples, right click select all rows shows incorrect dialog with number.
Guided workflow - even if required GO annotation columns are not present, wizard does not ignore GO analysis step.
PCA - chose interpretation I1 which has only one condition -> gives error -> choose interpretation I2 which has more than one condition -> works fine -> go back on the wizard and choose I1 again -> gives error -> but now can click Next and results from choosing I2 show up.
Clustering - Choose a linkage rule that results in out of memory error -> give proper error -> choose a linkage rule that works fine -> go back on the wizard and choose the linkage rule that doesn’t work -> gives error -> but now can click Next and old results show up.
Coloring of column names in hierarchical condition clustering is not proper in the cases where column name is different from sample name.
Description incorrect in p-value computation page for 2-way ANOVA in significance analysis wizard.
Iter-Plier crashes for 2 CEL files on Linux and Mac. It works fine for 1 CEL file or for 3 or more CEL files.
Project containing experiment created from custom Affy CDF when exported as project zip does not export the technology also.
After updating GeneSpring GX from version 9.0.2 to version 9.0.5, if you also purchase and upgrade to Workgroup mode, then connecting to the Workgroup Server using SSL will not work. The workaround is to connect without using SSL or to install version 9.0.5 of the client.

Note

This release is version 9.0.5 and follows public release 9.0.2. Intermediate releases 9.0.3 and 9.0.4 were never released publicly.

GeneSpring GX 9.0.2 - Released 7-Feb-2008

Last Update: 7-Apr-2008

Fixed Bugs

On Mac OS X, for non-US-English locales, the date field for the license validation was not being parsed properly. A patch for parsing this has been put in and GeneSpring will install and validate on Mac OS X with different locale settings.
In some non-US-English locales, float and decimal values are represented by a comma. These values were not being read appropriately, causing Fold Change Analysis in Advanced and Guided workflows to fail. This has been fixed and float and decimal values are now being read correctly in all locales.
Names in QC on Samples for Agilent arrays was scrambled. Fixed.

Known Issues

The following issues have been found in 9.0.2 and will be addressed in the next revision.

Tool crashes, when Hierarchical clustering algorithm runs with one sample, which is created by preprocessing with 'Baseline to median of all sample', in Mac. it works fine in Windows and Linux.
Bug in generic 1-color/2-color wizard - No proper error message if you choose wrong options in selecting rows. For example.. choose between marks gives blank error. eg2: if the column header is not selected, gives error "No categorical or continuous columns "
On correlation plot in QC on samples, right click select all rows shows incorrect dialog with number.
If you have two objects of the same name in a multi-inspector, on switching between the two, the inspector page for each does not change
Coloring of column names in hierarchical condition clustering is not proper in the cases where column name is different from sample name.
Guided workflow - even if required GO annotation columns are not present, wizard does not ignore GO analysis step.
Search - set of returned results from the enterprise is filtered out by the set of results that are not deleted. So reducing the max search limit will sometimes result in confusing results.
Condition tree - If you right click and open and if samples are removed, then it shows "Cannot launch" error followed by "0/-1" error. Double click open works as expected.
GSEA - only one of the two result tables "Gene Sets satisfying minimum genes" and "Gene Sets failing minimum genes" show.
GSEA - Clicking cancel while "Choosing Gene Sets with Minimum Matches" progress message gives two error dialogs in stead of one
GSEA - If you execute once, and then go back on the wizard and execute with a different set of inputs and cancel midways, old results are displayed.
GSEA - All genesets are converted to Gene Symbol before execution. If it is not possible to convert one of them, they are used as is. However their mark is different and only genesets of one mark are displayed in the results table. This happens in advanced search option only.
GSEA - every run of GSEA with same set of inputs can result in potentially different results. The random seed inside the algorithm is not fixed.
GSEA - use advanced search option -> choose lists -> run GSEA -> go back to choose lists page -> list empty.
PCA - chose interpretation I1 which has only one condition -> gives error -> choose interpretation I2 which has more than one condition -> works fine -> go back on the wizard and choose I1 again -> gives error -> but now can click Next and results from choosing I2 show up.
Clustering - Choose a linkage rule that results in out of memory error -> give proper error -> choose a linkage rule that works fine -> go back on the wizard and choose the linkage rule that doesn’t work -> gives error -> but now can click Next and old results show up.
Description incorrect in p-value computation page for 2-way ANOVA in significance analysis wizard.
Iter-Plier crashes for 2 CEL files on Linux and Mac. It works fine for 1 CEL file or for 3 or more CEL files.
Project containing experiment created from custom Affy CDF when exported as project zip does not export the technology also.
Tiff doesn’t show up in the drop down in the file chooser on export as image.
A condition tree built on average columns does not show the gene id column as labels.

GeneSpring GX 9.0.1 - Released 30-Jan-2008

Last Update: 30-Jan-2008

Fixed Bugs

When GeneSpring is installed in non-US-English locales, the date field was not being parsed properly and the tool would fail to create any new projects or experiments. Patch for parsing the date in different locale has been put in.
In generic single color and two color experiments, the "Export Entity List" functionality was not working. A Patch for "Export Entity List" functionality in generic one/two color experiments has been put in.
In generic one color experiments, the file chooser could only take in txt and csv files. XLS and gpr files, could not be used. In the current release, an "All Files" file filter in generic one color and "XLS" file filter in generic one/two color new experiment creation is available.
In the experiment creating wizard, for Agilent and two-dye experiments, the user interface for selecting dye-swapped samples did not have a scroll bar when there were many samples in the experiment. This has been fixed and a scroll bar appears when there are many samples in an experiment.
In Affymetrix experiments, the control probes for the Rice Affymetrix Gene Chip has been incorporated into the global Control probes. The quality plots for Affymetrix Rice Gene Chips will now show these control probes.

Known Issues

Excel files (.xls) should not be used to create a 1 or 2 color generic experiment, since samples created from these files give an error.
It is recommended that users that were affected by the non-US English locale bug should not update through the internal update mechanism (Tools -> Update Prodcut -> Update from the Web), but should uninstall (data WILL NOT be deleted) and install the 9.0.1 version. Updating through the internal update mechanism results in the demo project not being loaded.

GeneSpring GX 9.0.0 - Released 15-Jan-2008

Last Update: 20-Jan-2008

GeneSpring GX 9 is an exciting new chapter in the life of GeneSpring GX. It has been completely re-built on top of the Strand Life Sciences AVADIS development platform and provides unprecedented easy to use, yet, powerful gene expression analysis functionality. Listed below are a number of the new features found in this version.

New Features

Guided workflows for most common task

Task-specific guided workflows provide a clear path for the most common analysis, with instructions and data-dependent defaults, with a clear end point
Workflows can be specific for each vendor or application, allowing the user to make optimal use of vendor-provided QC
Advanced analysis also provides guidance, shortening learning curve

Project is the basis for all data organization

One window represents one project, that can contain all results, regardless of the particular chip the analysis was performed on

Hierarchy of analysis results maintained

Helps user track how particular analyses results were created by showing the results in a tree

Pathway visualization and analysis

Real interaction networks and pathways
BioPax exchange format for loading pathways and networks

Automatically translate a gene list from one technology to another

Supported for Human, Mouse and Rat

Support for all modern array vendors and formats

Agilent Feature Extraction files (Version 8.5 and later)
Affymetrix CEL,CHP and ARR files (AGCC and older binary and text formats)
Illumina Beadstudio (GeneSpring Export format, Probe Profile only)
Other formats can be imported as custom technology and GeneSpring can learn to recognize these new formats
Automatically update

New significance analysis algorithms (Parametric and non-parametric)

Three-way ANOVA
Unbalanced two-way ANOVA
Repeated measures
Paired T-test
Gene Set Enrichment Analysis (GSEA)

New normalization and baseline transformation options

PLIER
iterPLIER
MAS5
Li-Wong
Quantile Normalization

New Class Prediction algorithms

Decision Tree
Naïve Bayes
Neural Network

New Correlation measures

Squared Euclidean
Manhattan
Chebyshev
Differential
Pearson Absolute
Pearson Centered
Pearson Uncentered

New clustering algorithms

PCA-based clustering

New Visualizations

MVA Plot
Histogram
Matrix Plot
Vendor Specific QC plots

Automation

All functionality of GeneSpring GX 9 can be scripted with the JYTHON scripting language (PYTHON for JAVA)
Full integration of the R programming language

Supported platforms

Windows VISTA (32 and 64 bit)
Windows XP (32 bit)
MacOS X (10.4 and 10.5)
Linux (32 and 64 bit)

Known Issues

* Some functionality currently present in GeneSpring GX version 7.3, is not yet available in version 9.0 and will be considered for future implementation. A list of functionality not present in 9.0 is listed below

Quality Threshold clustering
Bar graph view
Ordered list view
Bookmarks
Find regulatory sequences
GROOVY Scripting language
Create gene list from annotations
3D scatterplots
Update annotations from Entrez Gene and other public databases
Update annotation from tab delimited file
Split window by Gene List folder

The following issues have been discovered with GeneSpring GX 9.0.0

IMPORTANT FOR MacOS X: On MacOS X 1.5 (Leopard), after running some features which use native code the program may get into an error state where running many of other tools. When it gets into the error state, the following error message is produced:

If this happens, unfortunately you need to re-install the application. This is a bug in Leopard and we have asked Apple for a solution to this issue.

After making the windows tile with "Both" to show all windows, selecting in a window will maximize that window, but only the first time. After you cascade again, you can select in the window.
Double clicking outside the pane for the condition tree will always open an entity inspector.
A classification profile plot will not show "split view" interpretations. It will only show unsplit Interpretations correctly. Split view interpretations will be shown as the "All Samples" interpretation
On some windows machines, license activation could fail with an error. The error will dialog with say Error 2501: Bad request. This happens because some run time dlls required form the VC++ 8.0 dlls are not found on some windows XP sytems. Even though we package it in the installer, the application is forced to pick it up from the WINDOWS/system32 library. These libraries are normally available in the system in the WinSxS directory. The solution is to install ./app/msvc_redist/vcredist_x64.exe (for 64 bit machines) and ./app/msvc_redist/vcredist_x86.exe (for 32 bit machines)
On rare occasions on Vista machines Derby will refuse to start up causing GeneSpring GX to crash. The stderr.log will have an error reporting lack of connection to port 1527. The solution is to run bin/packages/StartNetworkServer.bat for windows. Then launch GeneSpring GX again and it should start up.
P-Value’s that are exactly 0 on do not show up on the volcano plot. They do correctly show up on the table and the resulting entity lists contain the correct number of entities
The startNetworkServer.sh packaged with the application doesn’t work on MacOS X. It needs to be edited and "$PWD/../jre/bin/java" should be changed to just "java".
On some experiments, after migration when the box whisker comes up, it has a few annotation columns (eg, Gene symbols) which cause signals to become less visible. Close the box whisker plot and create new views and everything will be correct again.
A condition tree built on average columns does not show the gene id column as labels. Below find a script to show the labels correction. Replace the word 'Hierarchical' with the name of the condition tree view that needs to be changed. Also assuming that All entities and technology have same number of entities.

v = script.view.findView ('Hierarchical')
d = v.dataset
t = script.marray.project.getActiveExperiment().getTechnology()
t = script.marray.project.technology.createTechnology (t)
c = t.getAnnotationColumnByMark ('genesymbol')
d.addColumn (c)
v.labelByColumn = len(d) - 1

R scripts; running them when the R path is not set gives an exception in the script editor. The thing to do is to set the R path in Tools->Miscellaneous->R path; By default, the path is “C:\Program Files\R\R-2.6.1\bin”
Project zips for import have a 4GB limit; For larger projects, export zip will write out even when the size is larger than 4GB but the import will fail. There is no workaround known for this other than reducing the number of experiments being exported when exporting a project.
Tiff export for full dendrograms requires a small change (by default only png, jpeg and jpg appear in the file filter). To make tiff appear on the file chooser, delete C:\Program Files\Agilent\GeneSpringGX\bin\packages\pathway\viz\3.2\plugins\ImageFilters?.plg (keep a backup for safety). Now restart.
Search by default is restricted to 500 results (modifiable via Tools->Options->Misc->Search). A side-effect of this is that Find Similar Pathways will run only on the first 500 hits it gets. So it is essential to set the Options appropriately before doing Find SImilar Pathways.
On non-US or non-English Windows systems, the product may report the following error when creating a new Project:

“This project does not exist or you do not have permissions on this project”

To rectify this problem, replace the file “Startup.plg” in the following directory :

INSTALL_DIRECTORY/bin/packages/marray/product/1.4/plugins/

Migration from GeneSpring GX 7

Due to the completely new data organization in GeneSpring GX 9 when compared to previous versions, data that was saved in GeneSpring GX 7 is not automatically available in version 9, but migration tools are provided. There are two options for data migration, depending on need for the old data:

Re-create an experiment in 9 from the raw files in 7
Open the normalized data from 7 as a new experiment in 9

The first option will result in a new experiment in 9, with all the parameter settings etc. maintained, but the normalized values may not be exactly the same. This option is only available for experiments that were created with Agilent FE files and Affymetrix CEL and CHP files. The second option is available for all experiments from GeneSpring GX 7 and will result in an exact copy of the normalized values, but the samples from which the experiment was created are not available for the creation of a new experiment. In addition to the experiments, gene lists can also be migrated, but migration of trees and classifications is currently not supported, but these derived objects can easily be re-created in version 9. For option 1 (re-creation of an experiment), choose menu item Project -- > Import GS7 experiment and follow the instructions. For option 2 (Open normalized data from 7) the follow steps should be followed

Run the migration tool once

Menu item Tools --> Export GS7 Experiments
This tool only need to be run once
This tool will prepare GX 7 for migration

Create a custom technology

Menu item Tools --> Create Custom technology --> Import GS7 genome
This tool only needs to be run once for each Genome in GeneSpring GX 7
This tool will create a library for GX 9 that contains all the annotations for the genes

Import a GeneSpring GX 7 experiment

Menu item Project --> Import GS7 experiment
Choose the Export Normalized Experiment Signals