Canine Array FAQ

What’s the difference between v1 and v2?

Two array versions exist and have some different features. Version 1 has ~27,000 high quality SNPs tiled as PM/MM probes, making this a very robust and easy to use array for disease gene mapping. Version 2 has a larger number of high-quality SNPs (~50,000), but are tiled as PM probes only which means that a “feature effects” file is needed to run the genotyping algorithm. Alternatively, projects with a large number of unrelated, diverse samples (such that all genotypes are represented multiple times) can be analyzed without such a file. The version 2 array should be useful for trait mapping and also perhaps for identification of selective sweeps and copy number polymorphisms.

Version 1 details: The version 1 array (part number 520170) is a 5um, 100-format, PM/MM probe strategy (20 probe pairs/SNP), Whole Genome Sampling Assay (WGSA) design which can detect a total of 66K canine SNPs. These SNPs were chosen from the 2.5 million SNP map generated as part of the dog genome project. A “v1 platinum” set of 26,578 SNPs was selected to include accurate and robust SNPs, using a panel of >10 diverse breeds. The library file (DogSNPs520170P) has masked out the SNPs that are not of high quality and thus only will show results for the “v1 platinum” SNPs.

Version 2 details: The version 2 array (part number 520431) is a 5 um, 100-format, PM probe only (20 probes/SNP) WGSA design, which can detect a total of 127K SNPs. These SNPs were chosen from the 2.5 million SNP map generated as part of the dog genome project and include the majority of the v1 platinum set SNPs. A “v2 platinum” set of 49,663 SNPs was selected to include accurate and robust SNPs, using a panel of >10 diverse breeds. The library file (DogSNPs520431P) has masked out the SNPs that are not of high quality and thus only will show results for the “v2 platinum” SNPs. Also available is the library file for all SNPs.

How do I order arrays?

To order arrays and reagents please contact your local Affymetrix sales representative. To find your Affymetrix sales representatives please call the customer service number888-DNA-CHIP (888-362-2447).

  • version 1 custom canine SNP array (part number 520170)
  • version 2 custom canine SNP array (part number 520431)
  • GeneChip® Mapping 250K Sty Assay Kit (30 rxns) (part number 900765)
  • GeneChip® Mapping 250K Sty Assay Kit (100 rxns) (part number 900754)

Alternatively, the Broad Institute Center for Genotyping and Analysis can run your arrays as a service for a fee until Affymetrix is able to fully support this application.

What protocol do I use to run the arrays?

Both canine array versions are to be processed with the Mapping 500K Assay Manual (download from affymetrix.com). The canine array uses only the STY enzyme fraction, thus there is no need to prepare target for the other enzyme in the protocol. Additionally, the final volume of target DNA added to the array is 125 ul instead of 200 ul (due to the smaller array size). No other protocol changes are necessary.

It is recommended that laboratory personnel be trained on the human 500K assay before processing canine arrays.

Library files

Library files are needed for the Affymetrix scanner to interpret the probe pattern on the arrays. Library files need to be installed prior to running arrays. Please select the appropriate file from the below list:

 

Library File Name

To be used with Canine Array PN (version #)

 

#SNPs in library file

DogSNPs520170P

520170 (v1)

26,578 SNP v1 platinum set

DogSty06m520431P

520431 (v2)

49,663 SNP v2 platinum set

DogSty06m520431

520431 (v2)

127,132 SNP v2 full set

SNP identifiers and genome locations

To correlate the genotype information to the SNP names listed on the UCSC browserwe have made lookup files which translate the probe IDs to Broad SNP identifiersand CanFam2.0 chromosomal locations, and also provide allele conversion.

Analyzing v1 arrays using GTYPE

For projects with less than 40 unrelated samples use the DM algorithm to call genotypes. Please follow the instructions in the GTYPE manual located on Affymetrix.com. Note that you will receive an error message because the canine annotations are not present in NetAffx, click “ignore” to ignore the message. To download your genotypes – view the CHP results and then click on Export/Export Table in the top menu. This will save the genotypes and cel file names as a tab delimited file. Gender assignment is not possible since chromosome X SNPs are called in diploid mode.

For projects with more than 40 unrelated samples you can use the DM algorithm to perform a single array QC analysis. Then, once all samples for a project have been run, use the BRLMM batch analysis tool (following the instructions in the GTYPE manual located on Affymetrix.com. Please pay particular attention to the comments about number of samples needed and the reasons why.). Note that you will receive an error message because the canine annotations are not present in NetAffx, click “ignore” to ignore the message. To download your genotypes – view the CHP results and then click on Export/Export Table in the top menu. This will save the genotypes as a tab delimited file. Gender assignment is not possible since chromosome X SNPs are called in diploid mode.

Analyzing v2 arrays using the command line tool

Gender assignment is not possible since chromosome X SNPs are called in diploid mode.

Alternatively, the Broad Institute will help with calling genotypes until Affymetrix can support this software.

Using feature (probe) effect files

One feature effect file exists for each array version. They were generated using data from the same arrays used to select the platinum SNP sets and may improve genotype calling from your dataset. When using BRLMM or BRLMMp on the command line, the file is specified with the "--use-feat-eff" flag.

  • For version 1 arrays, the features effect file gives consistent calls with > 20 arrays from one breed.
  • For version 2 arrays, the features effect file gives consistent calls when run on a full plate of arrays. The plate can contain multiple breeds.

Chromosome X SNPs

SNPs on the X chromosome are NOT treated differently than SNPs on the autosomes. In other words, SNPs on the X chromosome will be called in diploid mode. Because of this, the genotyping software will not assign a gender.

Recommended QC procedure and experiment cut offs and accuracy

Version 1 arrays: We recommend only using arrays with a call rate of 75% or higher. Each call has p value indicating call reliability. Lowering the p value cutoff improves the call accuracy but decreases the call rate. We generally use p<0.25.

Version 2 arrays: We recommend only using arrays with a call rate of 75% or higher. Each call has p value indicating call reliability. Lowering the p value cutoff improves the call quality but decreases the call rate. We generally use p<0.01.

Scanner error message with version 2 arrays

If you are scanning v2 arrays and you get this error message: “Caught COM exception: Grid Align,” you need to install a patch on the scanner workstation. The patch and documentation is available here:

http://www.affymetrix.com/support/technical/software_patches/gcos_sp2_swupdate.affx

Broad institute project leader and contact: Dr Kerstin Lindblad-Toh (kersli@broadinstitute.org)