7. Interbull CoP - Data Exchange and time of evaluation

7.1 Data from participating organizations to the Interbull Centre:

7.1.1 Each participating organization is expected to provide:

A pedigree in a sire-dam format to be uploaded into the IDEA (Interbull Data Exchange Area) website,
Predicted genetic and genomic merit file for any trait the country participates with data in the international evaluations, and
File with parameters to be used in the international genetic evaluation.
1. All above files must be checked by the relevant checking program available inside IDEA
2. The .zip file produced for the pedigree and/or for the predicted geneti/genomic merit plus parameter needs to be uploaded into IDEA
Genotypes to be uploaded to the Interbull centre FTP server.

Countries are also expected to prepare files for genetic and genomic validation whenever (major) changes in national evaluations require that data is submitted for a test evaluation run, or when the previous genetic validation was done more than two years ago. Files should be prepared according to the applicable formats available in the Interbull webpage www.interbull.org

200 = pedigree file to be uploaded in IDEA;

300 = bull national predicted genetic merit file (all traits);

301 = parameter to be used in the international genetic evaluation (all traits)

Validation files:

302 = data file for conventional trendtest method 2

303 = data file for conventional trendtest method 3

300Df = data file containing daughter deviation file for GEBV test

300Cr = data file containing reduced conventional genetic evaluation for GEBV test

300Gr = data file containing reduced genomic evaluation for GEBV test

Conventional validation files, 302 and 303, should be used to run the conventional TrendTest software, genomic validation files (Df, Cr and Gr) should be used to run the GEBV test software. The zip file produced by the software should be sent back to Interbull Centre before the end of each test run evaluation.

7.1.2 Only pedigree information as present in the IDEA website is used for the international evaluation. After the deadline for submission of data has passed the uploading functionality in IDEA is blocked by Interbull Centre staff and the whole IDEA pedigree content is extracted. Breed-specific pedigree files, in a sire-dam format, are then reconstructed starting from all animals having a predicted genetic merit for any submitted traits.

7.1.5 Data for the international evaluations will normally be individual trait national predicted genetic/genomic merits (estimated breeding value, transmitting ability, or relative breeding value (RBV)) or SNP effects of bulls in the country’s most recent national evaluation. The predicted genetic/genomic merits should be provided in the same way as it is officially, and publicly, defined and expressed (reference base and official unit of measurement). All traits submitted to Interbull Centre for international evaluations must be described in detail using Form GE and Form GENO (described in Appendix II and downloadable from Interbull website www.interbull.org).

Exceptions from these requirements may be accepted as part of Interbull standard procedures for a trait-group and specified in the appropriate section of Chapter 6 “Traits and breeds”.

7.1.6 Countries submitting their data in RBV are particularly requested to include 2 decimal points in the variables, as explained in the file formats. De-regression and international evaluation will consider data in the original unit of measurement, i.e. not as RBV’s, but, as part of the service, international predicted genetic merits will be converted to RBV prior to release.

Countries submitting their data in RBV should provide separately the formulae and coefficients for back-transformation to estimated breeding values or transmitting abilities. Coefficients should be provided each time they are updated by submitting a new version of the relevant appendix in the Form GE (described in Appendix II and downloadable from Interbull website www.interbull.org).

7.1.7 Bull records from the following breed groups can be sent to the Interbull Centre: Brown Swiss-type, Guernsey-type, Holstein-Friesian-type, Jersey-type, Milking Shorthorn-type, Pinzgauer-type, Red Dairy Cattle-type (including several Red-and-White breeds), Simmental (including Montbeliarde)-type, and non-Red Dairy Cattle Red or Red-and-White-type breeds (e.g. MRY). Other breed groups will be added in the future if needed. Bulls should be classified under one of the above breed groups according to the definition given in each country, e.g. Red Holsteins could be categorized as Holstein-Friesian (if evaluated together with Black & White in a country) or non-Red Dairy Cattle Red depending on the direction the population has taken in this country. Individual countries should identify the breed groups their populations belong to. In case of cross-breeding, the breed with the highest percentage should be considered. International evaluations will be computed within the Brown Swiss, Guernsey, Holstein-Friesian, Jersey, Red Dairy Cattle, and Simmental breed groups, while other breed groups may be added in the future if needed and if investigations have shown the feasibility of corresponding international evaluations.

7.1.8 Data on all bulls (domestic and imported) evaluated in the participating country with daughters in at least 10 herds should be sent to the Interbull Centre. Individual countries are responsible for identifying the records properly with respect to type of predicted genetic/genomic merit, status of predicted genetic/genomic merit, status of bull, etc, as identified in the file format descriptions.

7.1.9 Data from each country are expected to be “clean” from foreign daughter information. Participants are required to address that situation in their submitted files.

7.1.10 Organizations are required to upload data (predicted genetic/genomic merit and parameter) into IDEA website (username and password are provided separately to organizations participating in the international evaluation). Genotypes need to be uploaded to the Interbull Centre FTP server.

7.1.11 Additional information like GE/GENO Forms should be uploaded into Interbull Centre FTP server (username and password are provided separately to organizations participating in the international evaluation) while percentage of red should be uploaded into IDEA using the AnimInfo module.

7.1.12 A detailed definition of the reference (genetic) base considered in the national genetic evaluation, as well as description of the base for age adjustment (if applicable), needs to be entered in the form GE/GENO (described in Appendix II and downloadable from Interbull website www.interbull.org) and accompany the data.

7.2 Data from the Interbull Centre to participating organizations:

7.2.1 One international predicted genetic merit file for each trait-group including bulls from all participating countries will be made available in each organization's own site on the Interbull Center FTP server together with information on the file format. The files will include bulls that are:

identified as officially publishable in at least one country;
bulls identified as part of a simultaneous progeny-testing program in at least 2 countries. International predicted genetic merits would be as defined in each country (estimated breeding value, transmitting ability, or relative breeding value) expressed in the relevant base and unit.

File formats are:

For Conventional MACE and GMACE:

030 = Interbull production predicted genetic merit file;

035 = Interbull conformation predicted genetic merit file;

135 = Interbull Brown Swiss conformation predicted genetic merit file;

036 =Interbull milk somatic cell and clinical mastitis predicted genetic merit file;

037 = Interbull longevity predicted genetic merit file;

038 = Interbull calving ease and stillbirth predicted genetic merit file;

039 = Interbull female fertility predicted genetic merit file.

040 = Interbull workability predicted genetic merit file.

For InterGenomics

InterGenomics files have no file code.

7.2.2 One file with international parent average reliability (ipr-files) and predicted parent averages for each trait-group including bulls from all participating organizations will be made available in each organization's own site on the Interbull FTP server together with information on the file format.

File formats are:

060 = Interbull production parent average reliability file;

065 = Interbull conformation parent average reliability file;

165 = Interbull Brown Swiss conformation parent average reliability file;

066 = Interbull milk somatic cell and clinical mastitis parent average reliability file;

067 = Interbull longevity parent average reliability file;

068 = Interbull calving ease and stillbirth parent average reliability file;

069 = Interbull female fertility parent average reliability file.

070 = Interbull workability parent average reliability file.

7.2.4 Each country providing data will receive evaluations for all breeds expressed on their base and scale.

7.2.5 Each country providing data will also receive evaluations for all bulls expressed on the base and scale of all other countries providing data. This information is not to be published by the country receiving it and may only be used for research and internal purposes.

7.2.6 A two-way table with conversion coefficients among all participating countries for all trait-groups evaluated, but subject to limitations identified in item 5.8, will be distributed.

7.2.7 Countries that do not participate with data, but wish to receive results from the international evaluation, will receive the same information as countries providing data. This information is not to be published by the country receiving it. It can only be used for research and internal purposes.

7.2.8 All participating organizations will have access to international pedigree files including all individual bulls. This information will be made available at the Interbull FTP server after each routine evaluation as one file per breed. The information in the pedigree file is strictly confidential and should be handled accordingly, and should only be used to prevent and to resolve identity problems.

7.2.9 Individual countries will be responsible for publishing results of the international evaluations according to own national requirements; publications should distinguish between national predicted genetic merits, international predicted genetic merits and national predicted genetic merits converted using the Interbull conversion factors.

All countries are expected to honor the Interbull “Guidelines for minimum requirements for advertising genetic merit of dairy animals” (Appendix V), and to enforce conformance with the Guidelines by all appropriate agencies at the national and international level.

7.3 Time of evaluations and deadlines:

7.3.1 Routine conventional and genomic international evaluations will be computed three times per year, in April, August and December. Test evaluation runs will be computed two times per year, in January and September.

7.3.2 Participating organizations should send their data for MACE routine evaluations to the IDEA database no later than by the Tuesday 21 days prior to a scheduled release of results. Submission of data for GMACE routine evaluations to IDEA database should be no later than by the Monday 15 days prior to a scheduled release of results. The definitive deadline for sending data is at 17:00 CET, but countries are encouraged to send data at earliest possible opportunity, since it would allow more time for checking and verification of the data, and communicating about possible inconsistencies. The most recent, correct, national evaluation file received by the deadline will be used in the international evaluations.

7.3.3 The target time for release of results for official publication is on the first Tuesday in April, the second Tuesday in August and the first Tuesday in December. The earliest possible official publication of results is at 07:00 local time, on the scheduled Tuesday. Target time for a confidential pre‑release of MACE evaluation results to countries having signed the “Interbull evaluation pre‑release agreement” (Appendix VI) is on the Thursday, 12:00 Central European Time (CET), 12 days prior to the scheduled release. Target time for a confidential pre‑release of GMACE evaluation results to countries having signed the “Interbull evaluation pre‑release agreement” (Appendix VI) is on the Tuesday, 12:00 Central European Time (CET), 7 days prior to the scheduled release. The pre‑release of evaluation results is strictly confidential and the sole intent is for checking of results, but it also provides the opportunity for countries to pre‑process the data prior to official release. The Interbull Centre may decide to release new evaluation results if a re-run is necessary, due to errors in procedures at the Centre or in national data discovered by the Centre or by a member after the pre-release. The Interbull Centre should send a report to all members on the likelihood that the Tuesday morning release will be possible to keep, no later than Friday, 17:00 CET.

7.3.4 Organisations participating to the IG-HOL evaluation should send their TMI data no later than by the Tuesday one (1) week after an official release of results. Target time for IG-HOL contributing countries to receive official TMI results is on the Tuesday, two (2) weeks after an official release of results.

7.3.5 Data submitted for test evaluation runs (items 4.1.2 and 4.2.4) should be sent by each new country, and by each country introducing changes in national evaluations, according to the following general schedule:

MACE TEST RUN: no later than 17:00 CET the Tuesday three (3) weeks after the release of August routine run and six (6) weeks after the release of December routine run. MACE test run results will be made available to countries four (4) weeks after.

GMACE TEST RUN: GEBV test results should be put on the Interbull FTP server no later than 17:00 CET the Tuesday two (2) weeks after MACE test run data reception deadline, GEBVs and file 734 should be sent to IDEA no later than 17:00 CET the Tuesday three (3) weeks after MACE test run data reception deadline. GMACE test run results will be distributed one (1) week after distribution of MACE test run results.

Truncated MACE: interested countries should submit their YYYY-4 data to the T-IDEA database no later than 17:00 CET the third Tuesday of October. Truncated Mace results will be distributed to participating countries three (3) weeks after.

Test results are confidential and are not meant to be published.

7.4 Test-run evaluation policy

7.4.1 Test-run evaluations intend to investigate the impact of new or modified national genetic evaluation results on the international evaluation; also to investigate potential improvements of the international genetic evaluation procedure. Results of such test-runs are reviewed by technical representatives of all countries involved. Pending on this review, changes are incorporated in the following routine evaluation.

7.4.2 Test-runs will be conducted in January and September at the Interbull Centre. The January test-run intends to investigate changes with the view to introduce them in the April routine evaluation of the same year. The September test-run intends to investigate changes with the view to introduce them in the December routine evaluation of the same year. More in general, partecipation in a test evaluation run with modified data implies that an official implementation of the modification is targeted for a near future, usually within 6 months.

7.4.3 Participating countries are requested to consider potential changes/modifications of their national genetic evaluation procedure in line with the above-mentioned time schedule.

7.4.4 Guidelines for assessing the necessity of a test-run:

The table below presents the actions required/taken during a test and a routine run for Mace and Gmace:

MACE Test run	1- Are any of the following rules applicable? (if not Go To 2) National EBV from a new country / breed / trait -> Go To 3 Two years since the last validation -> Verify program (check for consistency) -> Go To 2 National EBV from a new model -> Verify program (check for consistency) -> Go To 2 2- Check the consistency of the data by the “verify” program Large discrepancy -> Require explanation -> Accepted data -> Go To 3 Medium discrepancy -> Require explanation -> Accepted data -> Go To 3 Small discrepancy -> Accepted data -> Go To 3 3- Estimate the genetic variance using the REML method National data -> Estimate r_G -> MACE EBV -> Go To 4 4- Are the validation tests required? National EBV from a new country / breed / trait¹ -> Validation tests I-III (IV) -> Must pass Two years since the last validation -> Validation tests I-III (IV) -> Must pass National EBV from a new model ⁽²⁾ (if Δ_SDg> 5%) -> Validation tests I-III (IV) -> Must pass
MACE Routine run	5- Check the consistency of the data by the “verify” program National EBV from a validated model -> Verify program (check for consistency) -> Go To 6 6- Are the results of the “verify” program consistent with the results of the previous run? Large discrepancy -> Reject the data -> to be submitted to a test run -> Go To 1 Medium discrepancy -> Require explanation -> Reject/Accept the data -> Go To 1 or 7 Small discrepancy -> Accept data -> REML (check for consistency) -> Go To 7 7- Estimate the genetic variance using the REML method, and check for consistency (how large is Δ_SDg?) Large discrepancy -> Reject the data -> Revert to the last accepted data -> MACE EBV Medium discrepancy -> Require explanation -> Reject data -> Revert to the last accepted data ->MACE EBV Medium discrepancy -> Require explanation -> Accept data -> MACE EBV Small discrepancy -> Accept data -> MACE EBV
GMACE Test run	8- Check disparity between MACE and GMACE data National GEBV from a new country -> GEBV test -> Must pass-> Compulsory participation to a GMACE test run -> Go to 9 National GEBV based on the same national EBV as in MACE -> Go To 9 National GEBV based on different national EBV than MACE -> If no change in trait definition or genetic parameters -> Go To 9 National GEBV based on different national EBV than MACE -> If change in trait definition or genetic parameters -> Reject the data 9- Including data in GMACE Accept data -> GMACE
GMACE Routine run	10- Is any of the following rules applicable? (if not -> Go to 11) National GEBV from a new breed / trait -> GEBV test -> Must pass -> Go to11 Two years since the last GEBV-test -> GEBV test -> Must pass -> Go to 11 New GEBV model -> GEBV test -> Must pass -> Go to 11 11- Check the consistency of the data by the “Gverify” program National GEBV based on the same national EBV as in MACE -> Gverify (check for consistency) -> Go to 12 National GEBV based on different national EBV than MACE -> Reject the data 12- Including data in GMACE Large discrepancy -> A new GEBV test is provided -> Must pass -> Accept the data Large discrepancy -> No new GEBV test -> Reject the data -> Revert to the last accepted data -> Go to 11 Medium discrepancy -> Require explanation -> Reject/Accept the data Small discrepancy -> Accept the data-> GMACE

¹The rules for the traits to be validated are as before.
² If the Δ_SDg < 5%, the validation test is not needed.

7.4.4.1 Changes in models for genetic parameter and breeding value estimation and modification of pre-adjustment factors, such as age, require new validation according to the established procedure and a test-run before acceptance to a routine international evaluation.

7.4.4.2 Frequently countries introduce changes in their system with only minor effect on the national bull ranking. Such changes, however, may have considerable impact on international evaluations.One way to assess this would be to estimate the within-country sire variance with the procedure currently used by Interbull. Large changes of this parameter are invariably associated with substantial re-ranking in the international list.

7.4.4.3 Experience shows that changes in sire variance estimates (based on data included in the international evaluation) larger than 5% from one evaluation to the next can not be explained by added information from young bulls. In such case, validations and a test-run will be required.