"Transferring QTL technology to the pig breeding industry (PigQTech) a demonstration project"

Transferring QTL technology to the pig breeding industry (PigQTech) - a demonstration project

(Contract no. BIO4-CT97-2243)

Operative commencement date: October 1, 1997

Completion date: September 30, 2000

EC Contact: DGXII, Scientific Officer: Irene Norstedt, Freek Heidekamp

Costs: Total cost=2,024,235 ECU; EC contribution=1,220,000 ECU (60% of total cost)

Date of actual receipt of Community funds: October 1997

Project Coordinator: Prof. Leif Andersson, Dept of Animal Breeding and Genetics, Swedish University of Agricultural Sciences - SLU, Department of Animal Breeding and Genetics, Box 597, S-751 24 Uppsala, Sweden. Phone +46 18 471 4904; Fax: +46 18 471 4833; Email: leif.andersson@hgen.slu.se.

Partners:

Participant Nr 1:

Swedish University of Agricultural Sciences – SLU, Department of Animal Breeding and Genetics, Box 597, S-751 24 Uppsala, Sweden

Contractual status: Contractor, Coordinator

Participant Nr 2: Roslin Institute (Edinburgh) – RIO, Division of Genetics and Biometry, Roslin, EH25 9PS Midlothian, United Kingdom Contractual status: Associated Contractor

Participant Nr 3:

Pig Improvement Company – PIC, Fyfield Wick, Abingdon, OX13 5NA Oxfordshire, United Kingdom

Contractual status: Contractor

Participant Nr 4:

Scan Avel HB – SCAN, Box 505, S-244 24 Kaevlinge, Sweden

Contractual status: Associated Contractor

Participant Nr 5:

Cooperativa Agrícola y Ganadera de Lleida – COPAGA, Departamento de Producciòn Porcina, Polígono Industrial ”El Segre”, Avenida de la Indústria, P-203, Lleida 25080, Catalonia, Spain

Contractual status: Associated Contractor

Participant Nr 6:

Institut de Recerca i Tecnologia Agroalimentàries –IRTA, Area de Producción Animal. Centro UdL-IRTA, Alcalde Rovira Roure, 177, Lleida 25198, Catalonia, Spain

Contractual status: Associated Contractor

Participant Nr 7:

Universitat Autònoma de Barcelona – UAB, Dept. Patologia i Produccions Animals, Facultat de Veterinària, Bellaterra, 08193 Catalunya, Spain

Contractual status: Associated Contractor

A. Scientific achievements

Introduction

The aim of the project has been to demonstrate how the farm animal breeding industry can utilise gene mapping technology to accelerate genetic improvement. Previous theoretical studies had suggested that the use of marker assisted selection could potentially increase the annual improvement for quantitative traits like backfat with about 10% and for more difficult traits such as meat quality and reproduction by as much as 40-60% compared with existing technology.

The work has comprised two major tasks:

1. Commercially relevant populations have been screened for segregation at QTLs identified in experimental populations. The aim has been to establish optimal strategies for QTL detection in commercial pig populations and the extent to which QTLs explaining major phenotypic differences between divergent lines used in experimental studies also explain quantitative variation within commercial lines. The results are important for specifying future strategies for finding economically valuable QTLs.

2. Marker assisted backcrossing has been used to demonstrate how a QTL allele can be introgressed from one breed to another. The work has focused on the major fatness QTL on pig chromosome 4 previously identified in a wild pig/Large White intercross. The end result was not designed to be a commercially viable product in its own right, but the process has validated a number of points of major importance for the exploitation of QTLs in livestock.

Task 1: Establishment of database and network.

Participants: RIO (no. 2)

A project database was established which allowed storage, management and retrieval of project data. The project database was developed along with web based tools to access data. Data submission for the pedigrees and phenotypes was in the form of Excel formatted or text files sent by all participants to Participant 2 by email. Once these were loaded, participants defined markers and loaded marker genotypes into the database via their own web browser. Comprehensive error checking routines were built into the database and if errors were detected on submission of genotype data these were reported directly back to the user. This allowed individual users to check and correct their own data. The database also has a variety of tools allowing export of the data in different formats suitable for use with programs for linkage analysis, statistical analysis of phenotypes and QTL analysis. All phenotype and genotype data from pedigrees was loaded into the database.

A web site has been developed for the project outlining the purpose, participants, research, etc. At present this is private to participants with access via username and password, but once results are passed for general dissemination, these can be mounted on the site and the site made publicly accessible.

Task 2: Determine optimal design for detecting QTLs in commercial populations.

Participants: RIO (no. 2), IRTA (no. 6)

The objective was to optimise the sampling of animals to be genotyped so as to obtain the maximum information from a limited sample of all animals available. There are three major factors involved in optimising the sampling of a population in order to maximise the chance of detecting a QTL. These are the informativeness of the markers, the informativeness of the QTL and the pedigree structure. To increase the informativeness of the QTL, selection on the family variance proved effective, in that it reduced the number of individuals to be genotyped to detect a QTL of a given effect. The second major factor influencing power of QTL detection is the population structure. This boils down to the number of informative comparisons that can be made within a family, which is related to family size (number of full or half-sibs in a family). For example, a family of 8 full sibs gives 11 times the amount of information per genotype than a family of 2 full-sibs. Thus, large families are much preferred to small families. Similar arguments apply to half-sibs, i.e. large half-sib families are much preferred Our studies show that each half-sib pair of individuals is worth between 0.25 and 0.5 of a full sib pair. Thus, it is possible to get an approximate idea of the comparative value of pedigrees of different structure using these figures. We should also note that family size is a relatively robust measure of informativeness in a particular sample, i.e. it is valuable for both large and small QTL effects and is not biased in situations where selection based on the variance might be.

These considerations together with further calculations were used to select families from the target populations once Particpants 3 and 4 had summarised the families available for selection in terms of structure and phenotypic variation. In essence we focussed on large families (half-sibships) and within these, on families with the highest variance.

Task 3: Identification of candidate chromosome regions harbouring QTLs for growth and carcass traits.

Participants: All (1-7)

The QTL analysis focused on ten chromosomal regions (Table 1). Seven of these were candidate regions based on previous reports of QTL segregating in experimental populations while three were control regions for which no QTL had been reported at the start of the project. This design was chosen in order to evaluate to which extent QTLs segregating in experimental populations also is of major importance in commercial populations.

Table 1. Chromosomal regions selected for the PigQTech project.

Chromo-

some Interval Motivation Reference

1 CGA-Sw2185 Control -

2 IGF2-region QTL for % lean meat and back-fat thickness Jeon et al., 1999;

in WBxLW and LWxPietrain intercrosses Nezier et al., 1999

3 Sw72-Sw349 QTL for post-weaning Average Daily Gain Casas-Carrillo et

al., 1997

4 Sw35-Sw839 Major QTL for fatness and growth confirmed Andersson et al., 1994

in several populations

6 S0003-Sw316 Control -

7 MHC-region Major QTL for fatness in crosses between Rohrer et al. 1998

Meishan and European pigs

8 Sw905-Sw1029 QTL for carcass traits in a WBxLW Andersson-Eklund et al.,

intercrosses 1998

9 Sw983-Sw21 Control -

10 S0070-Sw1041 QTL for growth rate in a WBxLW cross Knott et al, 1998

13 S0068-Sw398 QTL affecting early growth rate (from birth Andersson et al., 1994

to 30 kg) in two independent studies Yu et al., 1995

LW=Large White; WB=Wild Boar

We decided during the project to include mitochondrial DNA (mtDNA) as an additional marker in the project. The reason for this was the finding by Participant 1 that pigs have been domesticated independently in Europe and Asia and that some major European breeds (Large White and Landrace) have a hybrid origin (Giuffra et al., 2000, Genetics 154:1785). The geographic origin of mitochondrial DNA is easily recognised by a DNA test. The animal material collected within the project was an excellent resource for testing the possibility that genetic variation in mtDNA affects phenotypic traits in pigs.

Task 4: Collecting phenotypic data and genomic DNA from target populations.

Participants: PIC (no. 3), SCAN (no. 4), Copaga (no.5), IRTA (no. 6)

Participants 3, 4, 5 and 6 collected about tissue samples and phenotypic data on about 5,000 pigs in total representing 10 different populations as outlined in Table 2. The selection of the material collected by Participants 3 and 4 followed the strategy established in Task 2. Pedigree records and phenotypic data were submitted to the database established by Participant 2 under Task 1. This strategy could not be used for the pedigree material collected by Participants 5 and 6 since the material were established already at the start of the project.

Tissue samples were transferred to genotyping laboratories for preparation of genomic DNA and marker analysis.

Table 2. Populations included in the project.

Breed Traits recorded

Contributed by Participant 3

Pietrain Growth, back-fat

Hampshire Growth, back-fat

Large White Growth, back-fat

Meishan X Large White Growth, back-fat

Contributed by Participant 4

SW Large White Growth, back-fat

SW Landrace Growth, back-fat

SW Hampshire Growth, back-fat

Contributed by Participant 5

Large White Growth, back-fat, meat quality

Pietrain Growth, back-fat, meat quality

Contributed by Participant 6

Landrace Growth, back-fat, meat quality

Task 5: Genotyping of target populations

Participants: SLU (no. 1) PIC (no. 3), UAB (no. 7)

Participants 1, 3 and 7 were responsible for genotyping the material as outlined in Table 3. This task included the preparation of genomic DNA and the analysis of at least one highly informative marker per selected chromosome region. Thus more than 50,000 genotypes have been generated during the course of this project. To facilitate this work some time were spent to select robust markers and to optimise conditions for multiplexing markers.

According to the original plan the genotyping task should be divided into two parts. The first one comprised screening one marker per chromosome region to be used for the detection of QTLs. This was accomplished according to the project plan. This activity was then planned to be followed by a second round of genotyping adding six markers to the most significant QTLs in order to provide better estimates of their location and effect. However, the delays encountered in collecting pedigree material as well as in generating genotype data did not allow the further screening of microsatellite markers (see further discussion below). However, this was compensated by typing mtDNA in all populations. This allowed us to make a large scale screening of the presence of Asian mtDNA types in European pigs as well as testing the possibility that variation in mtDNA influence production traits in pigs.

All genotype data were entered in the database established in Task 1 and were used for QTL analysis in Task 6.

Table 3. Summary of genotyping responsibilities

Material collected by Material genotyped by

Participant 3 Participant 3

Participant 4 Participant 1

Participant 5 Participant 7

Participant 6 Participant 7

Task 6: Statistical analysis

Participants: RIO (no. 2), IRTA (no. 6)

The objective of this task was to analyse the data generated within the project to demonstrate the presence of segregating QTLs in the populations sampled. We have currently analysed the data looking for QTL using the least-squares half-sib approach. We have also performed analyses of the effect of mitochondrial haplotype in all populations where there were sufficient numbers of each of the two haplotypes. The simpler, least squares based methods are rapid and robust and can be used to give a preliminary picture of the candidate QTL. Methods for performing this analysis simply require both phenotypic and genotypic data. During the course of the project we have developed a web-based software capable of performing least-squares half-sib analyses which was used for the analyses of the project data. This software, currently available at http://latte.cap.ed.ac.uk/ will allow other researcher or industry groups to readily perform similar analyses of their own data.

The results of the QTL analysis as regards growth and backfat thickness are compiled in Table 4 for the ten populations. We have carried out a total of 200 individual QTL tests (10 chromosomal regions x 10 populations x 2 traits). We thus expected to observe about 10 “significancies” at the nominal level (P<0.05) by chance only. However, we observed 22 significant effects or 26 if we take into account that four regions showed significant effects for both growth and fatness. We can thus safely conclude that we have observed more significant QTL than expected by chance. Some of the QTLs indicated in Table 4 are due to spurious effects, but many represent true effects of potential commercial significance. The results presented in Table 4 represent significant effects across all families for a particular chromosome and population. In addition, we observed individual sire families with suggestive QTL effects.

Table 4. Summary of QTL for backfat thickness (F) or growth/daily gain (G) detected in the PigQTech project. A significant QTL effect for the two traits in a particular population is indicated by an F or G, respectively.

Candidate regions Control regions

Population^a 1q^b 2^b 3 4 7 8 10 13 1p 6 9

Hampshire (3) - G G

Hampshire (4) - F F

Landrace (4) - F,G F

Landrace (6) - F F

Large White (3) G - F F

Large White (4) - F

Large White (5) - F,G F,G F

Pietrain (3) - F

Pietrain (5) - F,G G F

Meishan (3) F - G G

^a3, 4, 5, and 6 refer to which Participant provided the particular population

^bParticipant 3 decided to use a region on chromosome 1q instead of the region on chromosome 2 for commercial reasons. A dash in these two columns indicates that this region has not been tested in this particular population.

Several important conclusions can be drawn from these data:

1. A striking observation was the consistent QTL effect associated with chromosome 4. We observed significant effects in six out of ten populations. This result strongly suggests that Marker Assisted Selection using markers on chromosome 4 could improve commercial breeding programs.

2. Besides the consistent effect of chromosome 4, there were not significantly more positive test results for candidate regions than for control regions (Table 4). The results imply that it is recommended to rather investigate all parts of the genome in future QTL analysis of commercial populations. It is worth noticing that several recent reports have indicated the presence of a QTL for fat deposition traits in the region of chromosome 6 chosen as a control region on the basis of the information available when this project started. In fact, chromosome 6 was the region with the second largest number of positive test results in our study suggesting that this QTL is segregating at least in some commercial populations (Table 4).

3. The Meishan synthetic line showed positive QTL effects for growth or fat deposition on chromosomes 1q, 4 and 7. This result is entirely consistent with a number of reports describing the segregation of major QTLs in these regions in crosses between Meishan and European breeds. The results imply that Marker Assisted Selection could improve the efficiency in the construction of such synthetic lines.

4. There was no correlation between the presence of a particular QTL among different populations of the same breed. It is possible that the frequency of QTL alleles differs considerably between populations of the same breed but the result may at least partly be explained by a limited statistical power in the current study leading to a failure to detect segregating QTLs in some populations.

We have also analysed possible effects of the mtDNA type in those populations which contained sufficient numbers of the European and Asian type. The results are compiled in Table 5. The results show that there is no consistent effect of the mtDNA type on these two important production traits in pigs. The significant effect on growth observed for the Swedish Large White population may very well be a spurious effect due to chance or due to confounding with family effects.

Table 5. Results of association analysis between the mtDNA type and growth and fatness traits in PigQTech populations.

mtDNA type

Population^a Trait n European Asian Significance

Large White (4) Growth/day (gr) 444 559.4 595.3 0.01-0.001

Large White (4) Backfat (mm) 444 10.75 11.00 n.s.

Large White (5) Growth/day (gr) 188 571.0 562.3 n.s

Large White (5) Backfat (mm) 188 11.14 11.29 n.s

Landrace (4) Growth/day (gr) 478 579.8 540.5 >0.1

Landrace (4) Backfat (mm) 478 10.49 10.44 n.s.

Landrace (6) Growth/day (gr) 544 580.8 584.4 n.s.

Landrace (6) Backfat (mm) 544 13.30 13.51 n.s.

Pietrain (5) Growth/day (gr) 507 510.4 498.5 n.s.

Pietrain (5) Backfat (mm) 507 8.62 8.55 n.s.

^a3, 4, 5, and 6 refer to which Participant provided the particular population

Task 7: Introgression of a major QTL into a purebred line

Participants: SLU (no. 1)

This task involved the analysis of a major QTL on pig chromosome 4, first identified in a Wild Boar/Large White intercross. The QTL is located on pig chromosome 4 and has a large effect on fatness and growth. Interestingly, the results of Task 6 show that this is also a major QTL for fatness and growth in commercial populations. The aim of this task was to introgress the QTL allele from the Wild Boar associated with high fat deposition to a Large White population harbouring QTL alleles associated with leanness. A backcross pedigree (BC4) involving three recombinant sires were established with a total of 181 offspring, evenly distributed among the three sires. At least one QTL for fatness/meat quality was confirmed and the region was narrowed down to about 7 cM on the q arm of chromosome 4. One recombinant offspring from BC4 was selected and backcrossed to Large White animals to generate the 2^nd generation of the introgression experiment. The analysis of 50 progeny from this generation allowed us to formally exclude a part of the QTL region. The results of Task 7 demonstrate that it is feasible to introgress a QTL from one population to another and that the selective back-crossing can be used to improve the map localisation of a QTL. This is an important step towards the molecular characterisation of a major QTL and will facilitate the commercial exploitation of such a QTL since it is possible to use more closely linked markers for Marker Assisted Selection.

Task 8: Dissemination

Participants: All (no. 1-7)

As described in detail in the individual reports below, the scope of the PigQTech project has been widely disseminated to researchers and potential industrial partners within and outside EU. The PigQTech project is the most extensive project carried out so far to explore the usefulness of QTL technology applied to the pig breeding industry. Our presentations at international and national meetings have therefore raised considerable interest from the audience.

Concluding remarks

The PigQTech project has generated a considerable amount of important information as regards the application of QTL technology in animal production. We experienced some unexpected difficulties related to QTL analyses using material derived from commercial populations. One of these was that the project took place during a severe recession in the pig industry with falling prices. A consequence was that several pig producers rapidly decreased production, which made it much more difficult to collect the samples for the PigQTech project. This was the main reason why Participant 3 could not deliver samples from the target population on time. This in turn delayed the start of genotyping to the extent that Milestone 22 could not be completed during the project. Another problem with commercial populations was that we observed a very high rate of parentage errors in some populations. This also delayed the collection of the number of animals needed for the project. The observation implies that another benefit from a QTL program would be that parentage errors will be detected and thereby increase the precision in the breeding program.

The current project involved a large amount of marker analysis, more than 50,000 records. An important experience is that it is absolutely necessary to invest in good laboratory routines and bioinformatic resources in order to handle such large amount of information.

An additional achievement of the PigQTech project is that we have collected an excellent material for evaluating the importance of QTLs and candidate genes in commercial pig populations. This allowed us to test the possibility that variation in mitochondrial DNA may influence production traits in pigs. Our results did not indicate that this is the case. Similarly, participant 1 is currently using the PigQTech material to screen for genetic diversity in the KIT gene controlling dominant white colour in Large White and Landrace populations. We also think that the material now collected constitute a unique resource for testing the power of Linkage Disequilibrium mapping in commercial pig populations. The plan is to develop a high density marker map for some of the regions with confirmed QTLs and use these markers for Linkage Disequilibrium mapping in the PigQTech populations.

Individual reports

Participant no. 1:

Swedish University of Agricultural Sciences – SLU, Department of Animal Breeding and Genetics, Box 597, S-751 24 Uppsala, Sweden

Scientific Team: Dr. L. Andersson, Dr. V. Amarger, Dr. E. Giuffra, Dr. M. Moller, Mrs. U. Gustafsson, Mrs. S. Johansson

Task 3: Identification of candidate chromosome regions harbouring QTLs for growth and carcass traits.

The set of markers to be used in this project was established during the first year of the project. However, we decided during the project to include mitochondrial DNA (mtDNA) as an additional marker in the project. The reason for this was our observation that pigs have been domesticated independently in Europe and Asia and that some major European breeds (Large White and Landrace) have a hybrid origin (Giuffra et al., 2000, Genetics 154:1785). The geographic origin of mitochondrial DNA is easily recognised by a simple DNA test. The animal material collected within the project was an excellent resource for testing the possibility that genetic variation in mtDNA affects phenotypic traits in pigs.

Milestone 1: Candidate chromosome regions identified, complete.

Task 5: Genotyping of target populations

Tissue samples from the Hampshire, Large White and Landrace populations were received from Participant 4. Genomic DNA was prepared from all animals. Genetic markers representing the 10 selected chromosomal regions have been analysed. All genotypes have been interpreted, checked for pedigree errors and then submitted to the ResSpecies database (Participant 2) for QTL analysis. Since the sampling of the Large White and Landrace populations was so delayed it was not possible to genotype six additional microsatellites from QTL regions, so this activity was replaced by analysing the mtDNA type of all animals.

Milestone 10. Genomic DNA prepared for all animals, complete.

Milestone 17. Genotyping of two microsatellites per candidate region, complete.

Milestone 22. Genotyping of six additional microsatellites per population, not complete.

Task 7: Introgression of a major QTL into a purebred line

This task involved the introgression of a major QTL, first identified in a Wild Boar/Large White intercross, into a Large White population. The QTL is located on pig chromosome 4 and has a large effect on fatness and growth. A backcross pedigree (BC4) involving three recombinant sires were established with a total of 181 offspring, evenly distributed among the three sires. Fourteen genetic markers were used to determine the amount of the recombinant chromosome present and six markers were then selected for typing all animals for the QTL analyses. At least one QTL for fatness/meat quality was confirmed and the region was narrowed down to about 7 cM on the q arm of chromosome 4. One recombinant offspring from BC4 was selected and backcrossed to Large White animals to generate the 2^nd generation of the introgression experiment. The analysis of 50 progeny from this generation allowed us to formally exclude a part of the QTL region. The results of Task 7 demonstrate that it is feasible to introgress a QTL from one population to another and that the selective back-crossing can be used to narrow down the size of a QTL region considerably.

Milestone 3: Breeding animals for introgression experiment identified, 1^st generation, complete

Milestone 11: Records on fatness traits collected, 1^st generation, complete

Milestone 12: Analysis of chromosome 4 markers, 1^st generation, complete

Milestone 14: Breeding animals for introgression experiment identified, 2^nd generation, complete

Milestone 15: QTL analysis, 1^st generation, complete

Milestone 19: Records on fatness traits collected, 2^nd generation, complete

Milestone 20: Analysis of chromosome 4 markers, 2^nd generation, complete

Milestone 21: QTL analysis, 2^nd generation, complete

Task 8: Dissemination

The project was presented as an invited lecture “Marker assisted selection: issues and applications” during XXVI International conference on Animal Genetics, Auckland August 9-14, 1998

The coordinator (LA) was the local host for the EC Life Sciences Demonstration Conference held in Uppsala, October 11-14, 1998. The PigQTech project was presented as a poster and LA participated in a press conference in which some of the ongoing demonstration projects were presented.

Participant no. 2:

Roslin Institute (Edinburgh) - RIO

Roslin, Midlothian, EH25 9PS, United Kingdom

Scientific Team: Dr. A. Law, Mr. D. Nicholson, Dr. S. Blott, Dr. D. Chatziplis, Dr. G. Goosen, Mr. P. Shaw, Mr. G. Lee, Professor C.S. Haley

Task 1: Establishment of database and network.

A project database has been established which allows storage, management and retrieval of project data and which is accessible via the world wide web. The project database has been developed along with web based tools to access data. Data submission for the pedigrees and phenotypes is in the form of Excel formatted or text files sent to us by email. Once these are loaded, participants define markers and load marker genotypes into the database via their own web browser. There are comprehensive error checking routines built into the database and if errors are detected on submission or genotype data these are reported directly back to the user. This allows individual users to check and correct their own data. The database also has a variety of tools allowing export of the data in different formats. Thus data can be exported to check on the pedigree, genotypes or phenotypes. In addition data can be exported in formats suitable for use with programs for linkage analysis, statistical analysis of phenotypes and QTL analysis. All data from pedigrees and on phenotypes and genotypes from the project was loaded into the database.

The database and tools for loading and extracting data have in part been supported by the project and are now available and being used by other projects involving Roslin Institute and its collaborators. The development of the database has provided a striking demonstration of how critical good database/bioinformatics resources are to the success of a project such as this one.

Deliverables:

Milestone 4 : Achieved

Task 2: Determine optimal design for detecting QTLs.

The objective was to optimise the sampling of animals to be genotyped so as to obtain the maximum information from a limited sample of all animals available. We have spent some time exploring alternative sampling strategies using relatively simple pedigrees consisting of full-sib families nested within a half-sib structure. There are three major factors involved in optimising the sampling of a population in order to maximise the chance of detecting a QTL. These are: the informativeness of the markers, the informativeness of the QTL and the pedigree structure. To increase the informativeness of the QTL, selection on the family variance proved effective, in that it reduced the number of individuals to be genotyped to detect a QTL of a given effect.

The second major factor influencing power of QTL detection is the population structure. This boils down to the number of informative comparisons that can be made within a family which is related to family size (number of full or half-sibs in a family). For example, a family of 8 full sibs gives 11 times the amount of information per genotype than a family of 2 full-sibs. Thus large families are much preferred to small families. Similar arguments apply to half-sibs, i.e. large half-sib families are much preferred Our studies show that each half-sib pair of individuals is worth between 0.25 and 0.5 of a full sib pair. Thus it is possible to get an approximate idea of the comparative value of pedigrees of different structure using these figures. We should also note that family size is a relatively robust measure of informativeness in a particular sample, i.e. it is valuable for both large and small QTL effects and is not biased in situations where selection based on the variance might be.

These considerations together with further calculations were used to select families from the target populations once the companies had summarised the families available for selection in terms of structure and phenotypic variation. In essence we focussed on large families (half-sibships) and within these, on families with the highest variance.

Deliverables:

Milestone 2, 5 :Achieved

Deliverable 1 : Achieved

Task 6: Statistical analysis

The objective of this task is the analysis of data generated within the project to demonstrate the presence of segregating QTLs in the populations sampled. The simpler, least squares based methods are rapid and robust and can be used to give a preliminary picture of the candidate QTL. Methods for performing this analysis are in place and simply require both phenotypic and genotypic data. During the course of the project we have developed a web-based software capable of performing least-squares half-sib analyses which was used for the analyses of the project data. This software, currently available at http://latte.cap.ed.ac.uk/ will allow other researcher or industry groups to readily perform similar analyses of their own data.

The more complex methods can be very slow to compute and susceptible to non-normality or other problems in the data, but may give a more complete picture. We have developed a two step procedure to estimate QTL size and effect. Step one is where the probabilities of inheriting the same allele identical by decent (IBD) are calculated. Step two uses these probabilities, phenotypic and genotypic information in a maximum likelihood routine to determine putative QTL size and effect.

Additional data were generated in the project on mitochondrial haplotypes. These were classified into two main classes, European and Asian, and the effect of alternative haplotypes on performance was assessed. These analyses were performed in a mixed model context.

We have currently analysed the data looking for QTL using the least-squares half-sib approach. We have also analysed data from regions with an apparent significant effect in one population. We have also performed analyses of the effect of mitochondrial haplotype in all populations where there were sufficient numbers of each of the two haplotypes. Results of these analyses are given in the main body of the report.

Deliverables:

Milestones 13, 15, 18 Achieved

Milestone 23 Ongoing

Deliverable 5, 6: This report

Task 8: Dissemination

Haley, C., Kearsey, M., Knott, S., Seaton, G., George, A. & Visscher, P. (2000) User-friendly software to map QTLs in outbred populations. GAIT Workshop, Nottingham University, Nottingham, U.K., 12th-14th April, 2000.

Haley, C., George, A. & Knott, S. (2001) QTL mapping in livestock: progress and prospects. European Biotechnology Node for Interaction with China, (EBNIC), China/European Union Workshop, Shenzhen, People's Republic of China, 11th-13th October, 2000.

Participant no. 3:

Pig Improvement Company - PIC

Fyfield Wick, Abingdon, OX13 5NA Oxfordshire, United Kingdom

Scientific Team:

Dr. G.J. Evans, Dr. O.I. Southwood, Dr. T. Short, Dr. G.S. Plastow

Task 3: Identification of candidate chromosome regions harbouring QTLs for growth and carcass traits.

The objective of this task was to choose candidate chromosomal regions for screening of QTLs. Ten regions were chosen which comprised seven candidate QTL regions that had previously been reported in the literature and three control regions where no QTL had been previously reported. Dr Evans spent two and a half months from September-November 1997 in Uppsala, Sweden working with partner 1 on this task (see also task 5).

Milestones:

1. Identification of candidate chromosomal regions complete.

Task 4: Collecting phenotypic data and genomic DNA from target populations.

The objective of this task was to collect DNA and phenotypic information from sufficiently large numbers of animals from each pig population to allow a selection of the most informative 500 samples per population to be chosen. Four PIC populations were identified for use in the project, three pure breeds (Large White, Hampshire, Pietrain) and one recent Meishan synthetic line. The phenotypic information collected comprised growth rate and back-fat measurements. All phenotypic information was collected from the four populations and the records were submitted to the database established from task 1. This phenotypic information was used as model data by partner 2 to help to refine the optimal sampling strategy (task 2).

Using the optimal sampling strategy determined in task 2, the phenotypic variability within each litter was calculated. For each population the 10 sires with the most variable offspring were chosen. Then the largest, most variable litters were chosen from each of these sires until 50 offspring from each sire had been selected. This resulted in approximately 500 animals selected from each population for genotyping and should be the optimal selection of animals for detection of a QTL in the population.

Some initial delays were encountered in obtaining the phenotypic data which the delayed the use of the sampling regime to select the most informative samples for genotyping. Although this delayed the start of the genotyping, the collection of the data was timely enough for submission to the database (milestone 16).

Milestones:

6. Identification of sires to be sampled form target populations complete

7. Blood/Tissue samples collected from identified sires complete

8. Identification of animals to be sampled from the target populations complete

9. Blood or tissue samples collected from all animals complete

16. All phenotypic records submitted to database complete

Task 5: Genotyping of target populations

The objective of this task was to genotype each animal using two microsatellite markers per chromosomal region. Thus in each population, 2 markers per region x 10 regions x 500 animals = 10000 data points per population.

As mentioned Dr Evans spent a secondment period in Uppsala, Sweden in 1997 working with partner 1, investigating microsatellite markers from each of the candidate chromosomal regions (see also task 3). The markers were investigated for reaction conditions, robustness and ability to multiplex with other markers chosen for the project. Information on approximately five markers per region was initially gathered from which two per region were used to genotype the populations.

After selection of the animals to be genotyped from the target populations, DNA was prepared for all the sires dams and offspring. All sires were screened with the list of potential markers and multiplexes that were determined as mentioned above. The 20 most informative markers for each population were determined and routine genotyping was performed on the dams and offspring.

Routine genotyping began behind schedule due to the late data collection mentioned above. This was the first time that a project of such size and scope had been undertaken in this laboratory and delays were also encountered with generation and manipulation of genotypes. An upgrade to the PIC DNA sequencer was purchased in April 1999, helping to increase the generation of genotype data by 3 fold. However, manipulation of the generated data also became an issue. It is necessary to eliminate inheritance errors before genotypes can be accepted by the database developed by partner 2. Identification of these errors was extremely time consuming without an in house laboratory management database. Submission of genotypes to partner 2 was delayed due to these difficulties. It had been anticipated that there would be a need for a laboratory management database to manipulate the data. However, the only tool available specifically for this type of work was from an academic laboratory at University of Liege, Belgium. Despite working with this group in Cambridge and Liege it was not possible to obtain working version in time to be useful for the project. This system is still under development and no product support is available since it is not a commercial product. As this effort ultimately failed an alternative system was developed locally using a Microsoft Access database. This provided an immediate short-term solution that allowed the genotypes to be collected and successfully submitted to the database of Partner 2 for analysis. However, the delays encountered did not allow time for further screening of interesting QTL regions with extra microsatellites after the first analysis.

Deliverables

10. DNA prepared from all sires and offspring from four populations complete

17. Genotyping of two microsatellites per candidate region complete

22. Genotyping of six additional microsatellite per population incomplete

Task 8: Dissemination

Project introduced in “The role of major genes and DNA technology in selection for meat quality in pigs”, AG de Vries, A Sosnicki, JP Garnier & GS Plastow, Meat Science 49 Suppl 1 S245-S255 (1998).

Project described in “Advances in pig genomics and industry applications.”, Rothschild, M F and Plastow, G S. 1999 AgBiotechNet 10 p1-8.

7th Agrogene Seminar "Genomics and New Molecular Tools", Paris 25/26 Feb 1999.

British Pig Breeders Round Table (24-26 March 1999) Presentation - “Transferring QTL Technology to the Pig Breeding Industry - An EC demonstration project.

Project described in “The use of gene technology for optimal development of pork meat quality”. AG de Vries, L Faucitano, A Sosnicki, GS Plastow, Food Chemistry 69 397-405, 2000.

“Molecular genetics in the swine industry” (2000). Plastow G.S. Proceedings Anais Do III Simposio Nacional, Sociedade Brasileira de Melhoramento Animal. p21-30. Eds I J Nunes, F E Madalena & M A E Silva

Included in PIC Pork Chain Conference (Barcelona, Spain, March 2000): “Aplicacion de neuvas tecnologias para la selecion de carne de cerdo de calidad”

International Society of Animal Genetics, Minneapolis July 2000. PigQTech project discussed during a talk as an invited speaker.

Paper presented at China-EU Animal Biotechnology Workshop, Shenzhen, October 2000 (EBNIC). “Needs for Biotechnology in Animal Production”, G S Plastow, Conference Proceedings p131-140.

Participant no 4:

Scan Genetics, Sweden

Scientific team: M.Sc.Agric Ingela Hedebro Velander

Task 4: Collecting phenotypic data and genomic DNA from target populations.

Data structure and pedigree information of pure-bred pigs were investigated. In the data base of Scan all females and males have at least 3 generations of pedigree. That was enough for the project. An internal database for collecting data to PIGQTECH was evaluated.

Collecting genomic DNA from Scan Genetics three populations were done by sampling ear-notches from each pig in whole litters. About 25 herds were included from the start and ear notched were taken when technicians or the breeder himself were id-marking the piglets. The plan was to collect DNA from 20 litters of 20 AI-sires of Landrace, Yorkshire and Hampshire. The total aim of samples for the project was 3500 - 4000 pigs per breed. After collection period of ear notches and phenotypic data, 5 sires per breed with in total 500 offspring per breed were selected.

The ear notches was frozen and stored at Scan Genetics in Kävlinge. During this period technicians as well as breeders were involved. Visits to farms were made to check security in collecting DNA by ear-notches. The traits in question to be investigated were growth rate and back fat, measured at field testing. These data started to be available during february 1999. After ultrasonic testing a decision was made which of the sires that were going to be selected for the project. During this first collection part of the project there were some difficulties in collection of DNA from so many herds. The time it took to implement the new routine to all technicians, how to take samples and how to send the ear-notches from almost all litters in the herd, was longer than expected. The sampling itself was also very time-consuming at the farms.

All together during this project technicians have been sampling 8,200 ear notches (3080 H, 3100 L and 2020 Y) from offspring of selected AI-sires and pure bred dams in nucleus herds connected to the breeding program within Scan Genetics. A lot of people were involved in the collecting period and, at the same time, a decreasing pig market affected the nucleus herds. Two of the herds involved in the project from the beginning finished their production during this period. This meant that samples collected from these farms were not possible to use later on in the project due to missing phenotypic data. Also, for some sires just a few samples have been sent in due to the fact that some sires were not popular to use among the nucleus herds. The database for all collected samples, which was established the first period of the project, connected all phenotypic data with the samples taken. This lead to the decision of which H-, L- and Y -sires to use further on in the QTL-analysis phase.

During 1999-2000 phenotypic data were collected from each pig at about 100 kg live weight (range 80-130 kg). Back fat was measured by ultrasonic testing and at the same time the pigs were weighed by our technicians. Out of all collected data from the three populations 5 - 6 sires per breed and trait were selected for genotyping. This was done by analysing the variation of daily gain and backfat, between sires and the variation within litter for each sire. The aim was to select the sires with the largest variation due to these two traits to be further used for genotyping. Also, breeding values for each pig were evaluated. After analysis, data concerning 966 Hampshires, out of 8 sires and 173 dams, were delivered to the PIGQTECH database at Roslin institute. For Landrace the delivered data were 760 pigs out of 7 sires and 128 dams, and for Yorkshire 633 pigs, out of 6 sires and 115 dams.

The difficulties in this task were to find good routines for collecting DNA samples. This has not been done before in a larger scale. In this project we had to convince and teach farmers as well as our technicians to take the right samples among all other ordinary work in the herd. Therefor, the start up time was much longer than expected. There were a lot of samples collected that were not used in the analysis. This could have been avoided by a more precise plan from the beginning and number of sires or number of herds could have been minimised. On the other hand, collecting of DNA by ear notches is in the long run a good way of sampling in a normal nucleus production herd. There is a big difference in collecting in ordinary commercial herds than to collect from a research herd. Another unexpected problem was that not all animals were ultrasonic tested. This caused some problems and delay of the genome analysis as we had to wait for all the phenotypic values before we could select the animals for the project as only litters with at least 4 tested animals could be used. The project has demonstrated that it is possible and sometimes even recommendable to use common nucleus herds for collection of DNA on a larger scale as long as the choice of traits, herds, sires and animals are carefully specified and people involved get resources and proper instructions to do it. The accuracy in sampling has in total shown to be very high.

Deliverables:

Milestone 6: Identification of sires to be sampled from target populations, completed

Milestone 7: Blood or tissue samples collected from identified sires, completed

Milestone 8: Identification of animals to be sampled from target populations, completed

Milestone 9: Blood or tissue samples collected from all animals, completed

Milestone 16: All phenotypic records submitted to the database, completed

Task 8: Dissemination

Dissemination of the project has been made within the Scan company by one article in the internal monthly paper. At meetings with farmers and owners of the Scan company the project has been presented. Dissemination of the project has also been made at Pig breeders and producers conference in Sweden. The project has also been presented to the agricultural students in Alnarp.

Partcipant no. 5:

Cooperativa Agricola y Ganadera de Lleida (COPAGA)

Polígono Industrial "El Segre", 25191 Lleida, Spain

Scientific Team: A. Izquierdo, E. Gras, E. Ramells, L.Garcia, P. Borrás

TASK 4: Collecting phenotypica data and genomic DNA

Sires from Large White and Pietrain breeds and batches with offspring were chosen in collaboration with Partner 6. Copaga participated in the planning of appropriate matings and was responsible for animal management. All animals controlled descend from 5 sires in each breed. Blood was collected from all animals, sires, and dams and sent to Partner 7 for DNA extraction. Copaga provided the abattoir for slaughtering the animals. The traits recorded were:

- Weight at 120, 150,165 and 175 days of age

- Backfat thickness at 120, 150, 165 and 175 days of age

- Carcass weight

- Length of carcass

- Backfat thickness on carcass (two measurements)

- pH 45' and 24 h on longissimus dorsi and semimembranousus.

- Conductivity 45' and 24h on longissimus dorsi and semimembranousus

Meat quality analyses (fatty acid profiles and chemical composition) have been carried out in a subsample.

The total number of animals slaughtered and the family structure for the breeds analysed is in the Tables below.

PIETRAIN

Males

1 2 3 4 5 Total

Nº Sows 66

Nº Litters 20 19 12 12 11 74

Nº Offsprings 141 165 58 96 81 541

LARGE WHITE

Males

1 2 3 4 5 Total

Nº Sows 60

Nº Litters 20 14 9 13 19 75

Nº Offsprings 129 85 69 84 116 483

Task 8: Dissemination.

No scientific publications has been written yet, but two reports have appeared in two newspapers: El Segre (21^st january 1998, and Diario Cinco Días 13^th february 1998). The latter is a national information newspaper, and the former a local newspaper.

Participant no. 6:

Institut de Recerca i Tecnologia Agroalimentàries –IRTA

Area de Producción Animal, Centro UdL-IRTA, Av. Alcalde Rovira Roure, 177, 25198 Lleida, Spain

Scientific Team: J.L. Noguera, L Varona, J. Estany, D. Babot, M. Tor, D. Cubiló, M. Pérez-Enciso.

Task 2: Determine optimal design for detecting QTLs in commercial populations

The optimization of an experimental design in the Spanish populations accounted for multiple traits (meat and growth traits). The main parameter that can be changed is the number of sires. Based on simulations and on theoretical studies, we chosed five sires in order to reduce risk that all sires were homozygous for the QTL, and to increase power.

Task 4: Collecting phenotypic data and genomic DNA from target populations.

Five Landrace sires have been chosen and batches have been selected in a similar scheme as for Pietrain and Large White from Copaga. Blood has been extracted and sent to Partner 7. Copaga provided the abattoir for slaughtering the animals. The traits recorded were:

- Weight at 120, 150,165 and 175 days of age

- Backfat thickness at 120, 150, 165 and 175 days of age

- Carcass weight

- Length of carcass

- Backfat thickness on carcass (two measurements)

- pH 45' and 24 h on longissimus dorsi and semimembranousus.

- Conductivity 45' and 24h on longissimus dorsi and semimembranousus

Meat quality analyses (fatty acid profiles and chemical composition) have been carried out in a subsample. The total number of animals slaughtered and the family structure for the breeds analysed is given in the Table below.

LANDRACE

Males

1 2 3 4 5 Total

Nº Sows 80

Nº Litters 17 19 17 15 14 86

Nº Offsprings 116 109 119 102 105 551

Task 6: Statistical analysis.

We have developing a series of Statistical tools to analyse the data from this project using both Bayesian and classical approaches. We have focused our research on three main topics.

1. Computation of genetic relationships with DNA markers. The accurate estimation of the probability of identity by descent (IBD) at loci or genome positions of interest is paramount to the genetic study of quantitative traits. We present a Monte Carlo Markov Chain method to compute IBD probabilities between individuals conditional on DNA markers and on pedigree information. The IBDs can be obtained in a completely general pedigree at any genome position of interest, and all marker and pedigree information available is used. The method can be split into two steps at each iteration. First, phases are sampled using current genotypic configurations of relatives and second, crossover events are simulated conditional on phases. Internal track is kept of all founder origins and crossovers such that the IBD probabilities averaged over replicates are rapidly obtained. We illustrate the method with some examples. First, we show that all pedigree information should be used to obtain line origin probabilities in F2 crosses. Second, the distribution of genetic relationships between half and full sibs is analysed in both simulated data and in real data from a F2 cross in pigs.

2. Quantitative Trait Loci Mapping in F2 crosses between outbred lines. We develop a mixed-model approach for QTL analysis in crosses between outbred lines that allows for QTL segregation within lines as well as for differences in mean QTL effects between lines. We also propose a method called “segment mapping” that is based in partitioning the genome in a series of segments. The expected change in mean according to percentage of breed origin, together with the genetic variance associated with each segment, is estimated using maximum likelihood. The method also allows the estimation of differences in additive variances between the parental lines. Completely fixed random and mixed models together with segment mapping are compared via simulation. The segment mapping and mixed-model behaviours are similar to those in classical methods, either the fixed or random models, under simple genetic models (a single QTL with alternative alleles fixed in each line), whereas they provide less biased estimates and have higher power than fixed or random models in more complex situations, i.e. when the QTL are segregating within the parental lines. The segment mapping approach is particularly useful to determining which chromosome regions are likely to contain QTL when these are linked.

3. Computation of Bayes factors in QTL analysis. A fundamental issue in quantitative trait locus (QTL) mapping is to determine the plausibility of a QTL at a given genome location. The Bayesian analysis offers an attractive way of testing alternative models (here, QTL vs no QTL) via the Bayes factor. There have been several numerical approaches to compute the Bayes Factor, mostly based on Markov Chain Monte Carlo (MCMC), but these strategies are subject to numerical or stability problems. We propose a simple and stable approach to calculate the Bayes Factor between nested models. The procedure is based on a reparameterization of a variance component model in terms of intra-class correlation. The Bayes Factor can then be easily calculated from the output of a MCMC scheme by averaging conditional densities at the null intra-class correlation. We have studied the performance of the method using simulation. We have applied this approach to QTL analysis in an outbred population. We have also compared it with the Likelihood Ratio Test and we have analyzed its stability. Simulation results were very similar to the simulated parameters. The posterior probability of QTL model increases as the QTL effect does. The location of the QTL is also correctly obtained. The use of meta-analysis is suggested from the properties of the Bayes Factor.

4. We have reported the first preliminary results from the Spanish Pietrain population. Clear indications of the segregation of a couple of QTLs affecting several traits were obtained.

Task 8: Dissemination.

Communications of results of the project in several meetings has been presented :

Spanish Meeting of Quantitative and Applied Genetics (Caldes, Spain, June 2000). L. Varona, G. Davalos, M. Pérez-Enciso, J.M. Folch, N. Jiménez, A. Sanchez and J. L. Noguera. Análisis Bayesiano multivariante de componentes de varianza de caracteres de producción en un población Pietrain usando marcadores moleculares.

International Society of Animal Genetics (Minneapolis, USA, July, 2000). Varona, G. Davalos, M. Pérez-Enciso, J.M. Folch, N. Jiménez, A. Sánchez and J. L. Noguera. A Bayesian multivariate analysis with genome segment mapping: an application for production traits in a Pietrain population.

Participant no. 7:

Universitat Autònoma de Barcelona - UAB

Dept. Patologia i Produccions Animals, Facultat de Veterinària, 08193 Bellaterra, Spain

Scientific Team: Dr. A. Sánchez, Dr. J.M. Folch, Mr. G. Davalos, Mr. A. Clop, Mrs. N. Bello, Mrs. N. Jimenez.

Task 4: Collecting phenotypic data and genomic DNA from target populations.

The participation in this task has been to collect DNA from sufficiently large numbers of animals from three populations (Pietrain, Large White and Landrace) contributed by participants 5 and 6 to allow 500 samples per population to be analysed. The selection of samples has been carried out according to the optimal sampling strategy as defined from task 2. All DNA extractions have been performed according to data from animals slaughtered. A total of 1812 DNA samples have been extracted (Table 1). A duplicate of blood from each sample has been stored.

Deliverables:

Milestone 7. Collection of blood or tissue from sires, completed.

Milestone 9. Collection of blood or tissue samples in the three populations and DNA extractions, completed.

Milestone 10. Genomic DNA prepared for all animals, completed.

Task 5: Genotyping of target populations

The objective of this task was to genotype each animal in the three populations using informative microsatellite markers per chromosomal region. Microsatellite markers from each of the candidate chromosomal regions were analysed for optimal reaction conditions, robustness and ability to multiplex with other markers chosen for the project. Information on about five markers per region was initially provided. From these five markers, a minimum of two per region has been selected for genotyping the populations according their heterozygosity in the sires. Polymorphism of informative markers in sires was also analysed in dams for the three populations. The existence of at least one informative marker per each sire and chromosomal region was the basic selection criterion of markers. This goal was covered for most of the cases.

The total number of markers selected for analysis was 26 in the Pietrain population and 23 in the Landrace and Large White populations. After preliminary analysis of data, some evidence of a QTL segregating for several traits in chromosome 6 for the Pietrain population was obtained and 4 additional markers were analysed in this population. Genotyping of offspring for selected markers was completed for all 1812 samples received from the three populations. Four inter-laboratory control samples were genotyped for the 36 different microsatellites.

Microsatellite analysis of samples revealed that pedigree errors in the populations ranged between 2.7% in Landrace to 21.1% in Pietrain (Table 2). All animals of uncertain genealogy were discarded for statistical analysis.

Results were recorded and stored using the GEMMA software (D. Milan, INRA, France) and submitted to ResSpecies database (Roslin institute).

Deliverables:

Milestone 17. List of markers, PCR optimisation and multiplex conditions for each of the chromosomal regions completed. Genotyping of sires and dams for the three populations completed. Genotyping of markers for offspring in the three populations completed. Genotyping of additional markers for chromosome 6 in the Pietrain population completed.

Task 8: Dissemination

No scientific publications bearing specifically on this project has been written yet.

Reports and communications of the project during several meetings has been presented:

- Spanish Meeting of Animal Production (Zaragoza, Spain, May 1999).

- Spanish Meeting of Quantitative and Applied Genetics (Caldes, Spain, June 2000).

- International Society of Animal Genetics (Minneapolis, USA, July 2000).

Table 1. Samples received.

	Sires	Dams	Offspring	Total
PIETRAIN	5	66	549	620
LARGE WHITE	5	60	491	556
LANDRACE	5	81	550	636
TOTAL	15	207	1590	1812

Table 2. Samples of confirmed genealogy by microsatellite markers and used for statistical analysis.

	Sires	Dams Confirmed (total)	Offspring confirmed (total)	% pedigree errors
PIETRAIN	5	59 (66)	433 (549)	21.1
LARGE WHITE	5	54 (60)	388 (491)	20.9
LANDRACE	5	79 (81)	535 (550)	2.7
TOTAL	15	192 (207)	1356 (1590)	14.7

B. Table on objective fulfilment

Milestone (M)/Deliverable (D)	Task	Status	Partners
M1: Candidate chromosome regions identified	3	Complete	All
M2: Optimal sampling regime for two-generation families established	2	Complete	2,6
M3: Breeding animals for introgression experiment identified, 1st generation	7	Complete	1
M4: Database and network established	1	Complete	2
M5: Optimal sampling regime for three-generation families established	2	Complete	2,6
M6: Identification of sires to be sampled from target populations	4	Complete	3,4,5,6
M7: Blood or tissue samples collected from identified sires	4	Complete	3,4,5,6
M8: Identification of animals to be sampled from target populations	4	Complete	3,4,5,6
M9: Blood or tissue samples collected from all animals	4	Complete	3,4,5,6
M10: Genomic DNA prepared for all animals	5	Complete	1,3,7
M11: Records on fatness traits collected, 1st generation	7	Complete	1
M12: Analysis of chromosome 4 markers completed, 1st generation	7	Complete	1
M13: Methods for QTL detection in the target populations optimised	6	Complete	2,6
M14: Breeding animals for introgression experiment identified, 2nd generation	7	Complete	1
M15: QTL analysis completed, 1st generation	7	Complete	1
M16: All phenotypic records submitted to database	4	Complete	3,4,5,6
M17: Genotyping of two microsatellites per candidate region completed	5	Complete	1,3,7
M18: Preliminary QTL analyses of target populations completed	6	Complete	2,6
M19: Records on fatness traits collected, 2nd generation	7	Complete	1
M20: Analysis of chromosome 4 markers completed, 2nd generation	7	Complete	1
M21: QTL analysis completed, 2nd generation	7	Complete	1
M22: Genotyping of six additional microsatellites per population	5	Not complete	1,3,7
M23: Final QTL analysis	6	Complete	2,6
D1: Report on evaluation of optimal sampling regime for QTL detection	2	Complete	2,6
D2: Annual report.	-	Complete	All
D3: Evaluation of statistical methodologies for QTL detection	6	Complete	2,6
D4: Annual report	-	Complete	All
D5: Report on QTL detection in commercial pig populations	-	Complete	All
D6: Evaluation of the usefulness of experimental populations for identifying QTLs of commercial importance	-	Complete	All
D7: Evaluation of marker assisted backcrossing for the introgression of QTL alleles in commercial populations	7	Complete	1
D8: Refined mapping of the major fatness QTL on pig chromosome 4	7	Complete	1
D9: Evaluation of the prospects for positional cloning of major QTLs	7	Complete	1
D10: Route map for the commercial application of QTL research	-	Complete, attached	All
D11: Final report detailing progress in all areas of the project	-	Complete, attached	All
D12: Confidential report outlining exploitation plans	-	Complete, attached	All

C. Co-operation links

The strong links between the participants within this project are well illustrated by the summary of the project in section A (see Tables 2-5). In brief, Participants 3, 4, 5 and 6 collected tissue samples and phenotypic data from commercial pig populations and transferred these to genotyping laboratories and the PigQTech database, respectively. Participants 1, 3 and 7 were responsible for genotyping and transferred the data to the database. Participants 2 and 6 developed methods for the statistical analysis and were responsible for the statistical evaluations of the data generated by the other partners.

The project has been coordinated by regular meetings (two per year) and through frequent email contacts. Some exchange of researchers among the participants has taken place. Gary Evans (PIC) spent two and half months in the autumn of 1997 in Uppsala and worked together with Elisabetta Giuffra (SLU) to accomplish task 1. Alex Clop (UAB) spent altogether two and a half month in Uppsala, Sweden during the autumn of 1998 and in January 2000 to accomplish task 5.

The project has been presented to a wide audience as described in more detail in the individual reports under Task 8 Dissemination.

D. Illustrations

Good quality pictures of pigs representing some of the populations investigated in this study can be transferred preferably by ftp since the files are quite large.

(Does anyone have other pictures to provide?)

E. Scientific articles and books

Several additional papers are in preparation and will be submitted for publication during 2001.

Chatziplis, D. G., Hill, W. G. & Haley, C. S. (1998) Selective genotyping for QTL detection by sib-pair analysis. Proceedings of the 6th World Congress on Genetics Applied to Livestock Production, Armidale, N.S.W., 11-16 January, 1998. Vol. 26, pp 249-252.

Chatziplis, D. G., Hamann, H. & Haley, C. S. (2000) Selection and subsequent analysis of sib pair data for QTL detection. Genetical Research (submitted)

Chatziplis, D. G. & Haley, C. S. (2000) Selective genotyping for QTL detection using sib pair analysis in outbred populations with hierarchical structures. Genetics, Selection, Evolution (in press)

George, A. W., Visscher, P. M. & Haley, C. S. (2000) Mapping quantitative trait loci in complex pedigrees: A two step variance component approach. Genetics (in press)

Perez-Enciso, M. & Varona, L. 2000. Quantitative Trait Loci Mapping in F₂ Crosses Between Outbred Lines. Genetics 155: 391-405.

Perez-Enciso, M., Varona, L. & Rothschild, M. 2000. Computation of identity by descent probabilities conditional on DNA markers via a Monte Carlo Markov Chain method. Genet. Sel. Evol. 32: 467-482.

Varona, L, Garcia-Cortes, L.A., Perez-Enciso, M. 2001. Bayes Factor for detection of Quantitative Trait Loci. Genet. Sel. Evol. (in press).

F Press release

A good breeding programme is a prerequisite for an efficient animal production. For thousands of years animal breeding has involved the selection of breeding animals on the basis of the phenotypic characters of the animals. The rapid advance in the field of gene mapping technology holds the promise to improve these procedures by the possibility to identify animals with a favourable genetic constitution using DNA markers. In the EC-funded PigQTech project, four academic partners and three industrial partners from Sweden, Spain and the UK have now demonstrated how this technology can be applied to commercial pig populations. The consortium has sampled more than 5,000 pigs from 10 populations and screened these for 10 chromosomal regions of the pig genome. The project has demonstrated that this technology can be used to identify regions of the genome controlling important production traits such as growth and meat quality. We have also established routines for the collection of tissue samples from commercial populations and statistical methods for the analysis of data. The project paves the way for important commercial applications that will strengthen the competitiveness of the European animal industry. The technology can be used to improve production, fertility and animal health.

G. Acknowledgements

We thank the European Commission for the support to this project.