Software Support—Troubleshooting

The Mendelian error check analysis provides two key points of information:

• Are the input samples related (e.g., mother-child, father-child)?
• Do any chromosomes have an elevated occurrence of Mendelian errors?

The Mendelian error checking feature is for CytoScan™ arrays only.

The ChAS 3.1 browser allows for loading of different NetAffx™ versions at the same time (only if the versions are from all the same reference and genome builds).

The ChAS 3.1 browser cannot load NetAffx™ versions from different genome builds (Hg18 and Hg19) or from an external reference version (dbSNP build 135 and dbSNP build 132).

The algorithm is designed to detect only mosaicism between approximately 30–70% mosaicism for copy numbers between 1 and 3 for regions on the order of 5,000 markers in size or larger. The endpoint location of mosaic segments is less precise than the copy number segmentation, with endpoint variation within 500 markers being typical for segments of 5,000 markers or larger in size. Some regions of full integer copy number 1 or 3 below 5,000 markers in length may be incorrectly called as mosaic segments. Some regions of copy number below 1 or above 3, mosaic or otherwise, and less than 5,000 markers in length may be incorrectly called as mosaic segments.

No, it is solely a name change for consistency.

Yes, we have updated the cyto (CytoScan™ copy number) single-sample analysis engine with non- diploid analysis and correction. We have also converted the engine to make use of the APT2 analysis framework. The APT2 framework provides support for higher-precision internal calculations for the implementation of new algorithms, such as normal diploid correction, and for other future improvements.

The higher-precision calculations algorithm in CHAS 3.1 might result in small numerical differences when compared to previous versions of ChAS. The changes in results are smaller than those seen in technical replicates run through the same version of ChAS. For further details, please contact techsupport@thermofisher.com

While the reference model algorithm is unchanged, there are differences in file format used for reference model files. The reference model files provided have been updated to work with ChAS 3.1. Users generating their own reference model file will need to regenerate the model file in ChAS 3.1 to be able to use it.

Loading samples with a high number of segments may cause performance to be slower when publishing or promoting mosaic segments. Using “edit mode” and “fusing” fragmented segments before uploading is highly recommended.

No, once a file has been deleted, the only way to replace that file in the database is to publish the original xxchp file again.

In Control Panel \ Administrative Tools \ Service Check the chaspostgresql Service status is running If not attempt to start the service If it starts start the ChAS Database Service If it starts restart ChAS and see if the error is resolved. If the chaspostgresql service fails to start check the windows application event log. Look for a recent log file with a lock file "postmaster.pid (Ex. PID 2444) entry Start task manager and from the toolbar select view and select columns - ensure PID is ticked Locate the PID number found in the event log and right mouse click and select end task Go back to services and attempt to restart the chaspostgresql service then the ChAS Database Service.

Expression Console™ Software writes CHP files in GeneChip™ Command Console™ (AGCC) Software format by default. Some third-party analysis software packages may only be able to process CHP files in the legacy GeneChip™ Operating system (GCOS)-formatted CHP files. It is possible to save MAS5.0 CHP files in GCOS format using the ANALYSIS>Advanced Expression Configurations menu item. However, this is only available for CHP files analyzed with MAS5.0.

If metric and control information is not included in the header of the CHP file, Expression Console™ Software will not display that information. This occurs for two reasons: (1) an application other than Expression Console™ Software created the CHP file and did not write the information to the header or (2) the control probes were not identified prior to the analysis run and were not included in the header. To solve this issue in either case, rerun the analysis within Expression Console™ Software.

Two reasons explain why the numbers may be different between the two software applications. First, the development environment and compilers were different between the two software applications, resulting in minutely different behavior with floating point values. Secondly, the parameters used to call the algorithm may be different.

The Zip utility only includes library files for exon arrays when CHP files are analyzed as part of the study. If the user creates a study and adds existing CHP files, then the library files will not be included in the Zip archive. To add the appropriate library files, identify them through the CHP file properties in Expression Console™ Software and then manually add them to the Zip archive.

Examine the messages in the status window to identify the missing library files in Expression Console™ Software. If the computer is connected to the internet, use the download library files functionality to add the necessary files. In addition to these files, Expression Console™ Software requires two additional files for exonanalysis: the qcc file ith control probe information and an exon_analysis_configuration file. These files are automatically included when Expression Console™ Software is used to download the library files; however, if you manually add file to Expression Console™ Software, you will need to obtain the qcc and exon_analysis_configuration files from www.thermofisher.com.

You may benefit from using the SST-RMA when comparing data (significantly changed gene lists) with those processed using the RMA summarization only.

The mean of the background probe signal values is based on background probes defined in the background probe file, which are by default the anti-genomic probes. Anti-genomic probes consist of about 1,000 probes for each level of GC content (0 to 25) with no homology to most studied organisms. This set has a higher GC content than the average probe on the array, and therefore can have relatively higher signal values than the mean of all probes (PM_mean).

If a sample appears “on the line,” it is best to leave the sample in the experiment for analysis and simply flag it as questionable. During the initial analysis, treat all samples uniformly. Once candidate genes have been identified, review how the questionable sample changes data relative to the other replicates within the sample. It is always possible to remove a sample later in the analysis workflows.

Use either All_mean or PM mean to assay for hybridization intensity. All_mean is a probe-set metric. PM_mean is a probe-level metric, and is the mean of perfect match raw intensities prior to any transformations, such as normalization or probe summarization. PM_mean and All_mean can be compared to understand the effect that data processing steps have on the average intensity of an array because All_mean has been subject to any data transformations that have been performed during signal estimation and normalization. Apparent outliers only based on PM_mean can be ignored when corrected through data normalization in All_mean.

We do not recommend measuring the quality of a single-array performance without consideration for the rest of the experiment. In large-scale expression experiments using similar sample types, researchers are likely to develop their own single-array guidelines on what metric values are predictive of high- or poor-quality samples. However, these guidelines are likely to be dependent on sample type, and we are unable to recommend such guidelines for all possible situations. Note that the trend toward favoring model-based signal estimation algorithms (for all microarray experiments even beyond our platform) makes single-array quality determination very difficult due to the necessity of simultaneously analyzing multiple arrays to calculate signal estimates.

Pos_vs_Neg_AUC is a good first-pass metric. A value below 0.8 is a good indicator that sample problems exist. However, a value above 0.8 does not guarantee that the sample is good.

Sample variability, which arises mainly from biological heterogeneity, is certainly higher than assay variability, and has been estimated to be at least 10-fold greater. We recommend that researchers run multiple samples per data point to account for sample-to-sample variability. In addition, carefully design the experiment in order to minimize potential variation associated with the samples.

The list of array types for downloading library or annotation files in Expression Console™ or TAC covers the more commonly used array types. If you would like to analyze your expression data using our software, but your array is not on the list, please contact our technical support team for assistance. We can generate or provide the files necessary for analysis.

1. Check Operating System of computer and make sure compatible with the Expression Console™ version installing.
2. Check to make sure the library path for Expression Console™ is local and not under “My Documents”. In Expression Console™ software, Edit > Set library path
3. Uninstall Expression Console. Download the Expression Console™ package from the website. Unpackage the download. Right mouse click on the setup.exe file and “run as administrator”. Make sure the library path is local.

Check that the Netaffx™ password does not containing specific special characters # or & or + does not allow successful authentication. Change password online. Choose “My Account” Choose “Change Password” Update password that do not contain special characters. Try to log in to software again.

Genotyping Console™ Software cannot QC Mouse Diversity arrays. For QC one will need to use the Affymetrix™ power tools. There are Genotype and QC support files available on our website to help with this process.

Support for Affymetrix™ Power Tools (APT) is handled through the Affymetrix™ Developer Network. Questions, problems, feature requests and other issues can be addressed by emailing techsupport@thermofisher.com. Because APT is not a commercial product, support for this is not available via commercial product support channels.

Even though the browser being used is above IE version 7 this warning message can appear. To remove the message (Reference from AGCC 4.1.2 Release Notes) "If you log into the system and open the web portal under multiple user accounts, you may see a warning message indicating that Internet Explorer 11 is not a recommended browser. This message can be cleared by adding the local host website (the AGCC web portal) to the compatibility view list, by opening the Tools menu (click gear icon at upper right corner, or type ALT-X), click Compatibility View Settings, and then adding local host to the list using the 'add' button."

In addition to library and annotation files, the TAC software uses a configuration file that is normally obtained when you download the library and annotation files within TAC (termed Array Type files in the TAC Preferences menu). To run your analysis, download these files from within TAC, under the Preferences menu. Alternatively, if your array type is not listed or your analysis workstation is not networked, please contact our technical support team and we can generate or provide the TAC configuration file needed.

In the TAC software, you can filter the list of results by a list of gene symbols. This will also limit the hierarchical clustering of your expression data to the gene symbols found on your list. To do this, proceed with your analysis as you normally would. Once the results are returned, in the Table tab, you right click on the Gene Symbol column and choose the Filter option. A pop up will then appear allowing you to apply the filter logic. To filter on a list of Gene Symbols, select either the ‘In List’ or ‘Not In List’ option, as it pertains to your query. The Edit List window will appear, allowing you to type or paste in gene symbols. These must be separated by comma, semi-colon, or line breaks to be recognized properly by the software. Please also note that we use the official gene symbol determined by the NCBI Gene database. If your gene symbol of interest does not appear to function properly, please confirm that it is the official gene symbol by consulting the NCBI Gene database website.

If a sample appears “on the line”, it is best to leave the sample in the experiment for analysis and simply flag it as questionable. During the initial analysis, treat all samples uniformly. Once candidate genes have been identified, review how the questionable sample changes data relative to the other replicates within the sample. It is always possible to remove a sample later in the analysis workflows.

When you install catalogue array library files three threshold files (Diploid.ac_thresholds, Human.ac_thresholds, Polyploid.ac_thresholds) will be installed as well, they are common for all arrays hence they are automatically put in the main library folder not in a specific array subfolder. These three threshold files do not come with custom array library files though, so before doing a custom array analysis you install a catalogue array library file set first.

In order to export in VCF format from the Axiom Analysis Suite, the annotation files need to have the ref and alt allele columns. If these columns are blank in the annotations, the software will export "./." in place of genotypes.