Wednesday, February 29, 2012

PA, AA and Your Mudi

Today I uploaded the necessary information for you to be able to order PA. The info is on the Pedigree and Ancestor Analysis page of this blog.  It is written in English on that page, but it is currently available in Hungarian via download and also English via download (by links at the top of the page).  German download will be available soon.

A new page has been added to the blog: PA Certificate List

COI requests are always available at no cost (as in FREE), simply by sending me an email request to: mudicompass@gmail.com

Several of the other pages on this blog have had their information updated as well.  So as of today, this blog is up to speed!

As always, if you have questions or comments, you are welcome to email me!  gonduzomudi@gmail.com

If you want to know more about PA (pedigree analysis) and/or see the epilepsy statistics in the Mudi breed (Nov 2011), please see the January 2, 2012 entry of this blog.  (Updated statistics are coming soon)




Monday, January 2, 2012

Pedigree Analysis, Epilepsy and the Mudi


In these past few months I have been working on a very lengthy project.  It began with furthering my knowledge about Pedigree Analysis and then with the writing of an article which would explain as simply as possible what Pedigree Analysis is all about. The first article was then followed up with a statistical data collection and analysis report involving epilepsy in the Mudi. I asked two authorities on the subject to review my first and main article, Dr. Jerome Bell and CA Sharp.  Both kindly accepted and I incorporated their recommendations into the final draft of the article. I also asked Dr. Péter Pongrácz to review the articles and to advise and assist me with the statistical data procedures as that is one of his many areas of expertise.  I asked several Mudi owners to proof read the article as well as offer their insight (Michelle Murvai, Sharon Burkhardt, Betty Lessard).  And finally I asked Péter and Dörte Kolkmeyer to translate the articles to Hungarian and German as they are native speakers.  

I was very fortunate to have such willing and competent advisers and assistants to see me to the end of this part of the project.  This project and a COI project will also be shared with DogHeirs and the COI Source in the near future.  

All of those involved hope you will see the benefit of this work to the Mudi breed as well as to your own dogs and breeding plans.  

This Pedigree Analysis project initially consists of a 3 part series, more parts may be added later.  The first two parts are published below in English, but all 3 are downloadable by clicking on the name of each article below.  I have decided to publish on this blog the 2nd article first, the 1st article after it and the 3rd article will be placed to the PA tab on this blog shortly.  If you have any problems downloading the articles, email me and I will forward them to you.

I am currently working on the Mudi breed hip dysplasia statistics report and hope to publish that information very soon.

Deutsch
Pedigree Analyse – was ist das?.pdf
Epilepsie beim Mudi.pdf

Magyar
A pedigréanalízisről egyszerűen.pdf
Epilepszia a mudi fajtában.pdf

English
Demystifying Pedigree Analysis.pdf
Epilepsy in the Mudi.pdf



Epilepsy in the Mudi
Breed Statistics as of November 2011
Celeste R. Pongrácz

Genetic epilepsy has occurred in the Mudi breed since at least the 1980’s.  During the latter part of the last decade, epilepsy was not widespread and most people involved in the breed in that era did not share the information amongst themselves or with others.  We are reaping the fallout from that negligence today.  Since the autumn of 2008, it is impossible to ignore the fact that genetic epilepsy is the Mudi breeds number one health issue.

In order to attempt to lessen the occurrence of genetic epilepsy in the Mudi, it is first necessary to know what the breed average risk score is according to Pedigree Analysis.  With this risk factor we can plan better litters to bring the risk levels down and thereby reduce the occurrence of new epileptics.  If we do not reduce the breed average level significantly over the next few years, we will continue to see an exponential appearance of new epileptic Mudis. 

During the month of November 2011, 145 litters were scored using Pedigree Analysis.  The 5 generation pedigree of each litter was used to obtain the breed average epilepsy risk.  All Mudi litters born in 2008 and 2009 that were entered in my database were used to reach the score (145 litters: 73-2008; 72-2009).  These litters were born in 9 countries:  Hungary (115); Finland (10); Sweden (6); Germany (4); USA (3); Austria (2); Belgium (2); Czech Republic (2); Holland (1).

The epilepsy carrier and suspect carrier potential in my database are based on 18 known epileptics.  These 18 known epileptics were born between 1982 and 2008 as listed below:
1980’s – 4
1990’s – 2
2000 – 2
2002 – 2
2004 – 1
2005 – 1
2006 – 1
2007 – 3
2008 – 2
There are also 11 possible epileptics, born since 2002, that have not been confirmed yet, data from these 11 was not used to reach the average risk score.  There are also “Mudi’s” that have been placed via rescue groups in Hungary that have also been reported to be having seizures in their new homes, but these are also not included in this research.  It is essential to realize that epilepsy is not just found in the pedigreed Mudi population, but also in the not-pedigreed Mudis, Mudi crossbreeds, and the current and potential B/R registered Mudis.  There are many breeders in Hungary who have introducing B/R Mudis for breeding in this past year, however, eliminating the pedigree will not eliminate the risk, in fact it could do just the opposite.

The COI of the individual litters did not have a significant correlation to the highest risk scores.  The litters with the highest epilepsy risk scores were nearly equally divided above and below the recommended 12.5% maximum COI level (14 litters had an epilepsy risk score between 75 – 150; 8 had a COI higher than 12.5%).  The average COI score for all the litters born in 2008 was 9.6% and for 2009 it was 10.6%.  The high/low COI for 2008 was 22.6%/0.0%. The high/low COI for 2009 was 33.1%/0.0%.  (The COI calculation is run on the maximum number of known generations for each litter, for some litters this is as many as 25 gens, but the COI currently stabilizes around 10 gens.)

LITTER/YEAR SCORES:
Risk Score Definition:
Very Low = 0-12
Low = 13-24
Moderate = 25-49
High = 50-74
Very High = 75-99
Extreme = 100-200

2008 high/low score: 150/2 (2008 average: 45)
2009 high/low score: 109/3 (2009 average: 42)

Very High to Extreme Epilepsy Risk Score
- 6 litters scored 100 – 150
- 8 litters scored 75 - 99

High Epilepsy Risk Score
- 40 litters scored 50 - 74

54 litters, which are 37% of the litters born in 2008/2009, had a high to extreme epilepsy risk score!  That’s more than 1/3 of the litters born in 2 years!

Moderate Epilepsy Risk Score
- 58 litters scored 25 - 49 (58 litters = 40%)

Very Low/Low Epilepsy Risk Score
- 33 litters scored 2 - 24 (33 litters = 23%)

Every litter had a score above 0.  In other words, there was not one litter out of the 145 born that had no connection to epilepsy on their 5 generation pedigree.

The Mudi breed average, based on these 145 litters, is 44.  For comparison, other breed averages for epilepsy are: Australian Shepherd - 53; North American Shepherd (aka Miniature Australian Shepherd) - 40.

If you take into consideration that all of the epileptic Mudis that have occurred are not known, the actual Mudi breed average is probably higher than 44.

The breed average epilepsy risk scores for those countries that had 3 or more litters used in this study are:
Hungary - 43
Finland - 47
Sweden - 46
Germany - 60
USA - 47

4 of the 5 countries that produced the most litters in these 2 years, scored above the breed average!

SUMMARY
The facts are alarming:
1) The Mudi breed average epilepsy risk is 44.
2) The vast majority of litters (77%) are in the moderate to extreme risk score range.
3) There was not one litter without some risk.
4) Many of the countries producing Mudi litters are creating litters above the breed average.  
We need to alleviate the serious threat that epilepsy poses to the Mudi breeds future health and well being if the breed is to survive into the next decade.  Every litter needs to be planned with the utmost care in regard to the epilepsy risk factor associated with it. 

Avoiding epilepsy must become the priority of every Mudi breeder.

Future:  I plan to do this same research on the litters born in 2010 and 2011 in 2012.

(Copyright 2011, Pongrácz.  Permission: Translation and reprinting are allowed as long as the article is used in its entirety.  This article is best used in conjunction with the companion article “Demystifying Pedigree Analysis” written in November 2011 by Pongrácz.)

I would like to acknowledge and thank my reviewers and translators,
Michelle Murvai, Dr. Péter Pongrácz (Hungarian) and Dörte Kolkmeyer (German).  Their continued contribution to this research project is immeasurable!

_________________________________________________

Demystifying Pedigree Analysis
Celeste R. Pongrácz

Introduction
Basic Information
Cause and Effect (Why diseases occur and avoid control)
Breeding Options (Available choices)
Coming to Terms (A few words and abbreviations you need to know)
Data Collection (The how, where and disadvantages of data collecting)
Know Your Enemy (What traits to evaluate with PA)
Goals and Methods (What PA can do; Comparison of the two currently used methods)
Interpreting Scores the Sharp Way (More Sharp method score details)
COI Goals (What should be accomplished with the COI score)
Damage Control (Steps that can halt the loss of genetic diversity, the advancement of diseases and still allow the creation of a sound dog)
Take Home Summary (The important information to consider)
Links
References

Introduction
What is Pedigree Analysis?  Why do I need it?  Do I need to know genetics and scientific terminology to be able to understand it and use it?  So many questions and sometimes the answers given are not consistent.  When something goes over our heads we tend to avoid it. But, avoidance behavior can have serious consequences.

The purpose of this article is to explain what pedigree analysis is and what are the benefits of using it. I have written it mostly for the breed I am involved in, the Mudi, but this information can be applied to practically every breed.  Pedigree analysis (PA) is an evaluation of known relationship data, processed with mathematical formulas. It is a straight forward and simple process based on a standard set of rules.  It is also called relative risk assessment or relative risk pedigree analysis. To benefit from this process, you will need to familiarize yourself with a few terms which cannot be avoided, and you should know what a five generation pedigree looks like.

Some of the information I cover in this article does not pertain solely to PA, but in order to understand what can be accomplished with PA, it is necessary to know common breeding principles and their effects.

Basic Information
Genes carry the information to create every dog’s traits, characteristics, qualities and diseases.  However a dog is more than just the product of its genes, as interactions between genes and environmental agents, as well as learned behavior, also play an important part in the final composition of every dog. For PA purposes, the genes your dog inherits from its ancestors are what we are most interested in. Basically speaking, half of your dog’s genes are given from its mother and half from its father.  Every breed has certain “fixed” traits, that is, they have a typical size range, a typical ear habit (hanging, pricked, etc.), a typical color (white, black, brown, etc.), pattern range (merle, brindle, black & tan/bi-color, etc.), a typical purpose (companion, herding, hunting, etc.), typical fur type and so on.  Every breed has a basic trait list of what makes it one breed and not another.  The genes which make these basic traits become consistent because their variations are eliminated, that is, the genetic choices become invariable so the same features can be duplicated over and over.   

However, there are breed features which require several gene sets to appear, making them variable beyond the basic setting to some degree.   This is what breeders most often concentrate on changing, based on the current preference or need.   Knowing what traits the dogs listed on a pedigree had, or did not have, can be useful in this endeavor. 

PA is looking at a pedigree and trying to determine what kinds of specific genes the dogs on that pedigree may contribute to the dog the pedigree is for. In other words, Spot’s pedigree you now hold in your hands can show you the possible genes his ancestors may have given to him and the genes he may have to offer his pups.  These genes can be good, bad or neutral.  Every dog has a given set of genes and you cannot change them in the dog you have in front of you, only in the dogs that are yet unborn.  PA can be used to select one gene variant over another which can affect the conformation, performance ability and health traits of future pups.  

Responsible breeding requires maintaining a harmonized process of gene exchange and preservation. This involves persistent, broad based monitoring and balancing of the available gene set. Unfortunately breeding has mostly been done through a visible, results based screening and selection process, but this method does not provide the desired outcome consistently. This is why deleterious issues continue to emerge and increase within breeds and persistently uniform offspring cannot be achieved.

PA is not the perfect solution for breed health, performance or appearance problems. However, it can show you options to possibly keep or exchange one gene for another in Spot’s future litters.  It is far from an exact science to accomplish this.  If it were, we would not have the many problems dog breeds face today.  The keeping or exchanging of genes usually does not happen in one litter, it most often takes several generations after Spot to bring in what you want or to replace what you don’t.

Breeding healthy, temperamentally stable and performance able dogs today is proving to be difficult in spite of all the advancements that science and technology have made. The traditional breeding practices and choices that have been promoted for decades are not useful for avoiding diseases or promoting genetic diversity. Inbreeding is most often blamed for the loss of diversity in a breed, but it is the effect of artificial selection and popularity that are the main causes of genetic diversity loss.  The over use of popular sires and bitches, popular kennels, popular show and sport dogs and other “personal” selection criteria for or against certain dogs or lines, affects gene pool diversity more than high levels of inbreeding. The obsession with winning dog shows or performance events most often drives the need to breed only the best to the best, which means the best become the enemy of the rest!  Breeding generation after generation for extremes removes the majority of the population from breeding.  Breeding for constant “improvement” appears to have taken precedence over breeding for overall fitness.

Inbreeding and line breeding still have their place in today’s breed problems however.  These well known methods are used to ‘fix’ genetic traits into a line, but it is not a simple choice between ‘fixing’ only the traits you want from those you don’t. You cannot pick and choose genes as if they were on a restaurant menu. Inbreeding does not create desirable or undesirable genes, it simply gives a higher chance that a gene will be able to show its effect by making it more common.  This effect can be good, neutral or bad. You cannot get rid of a gene.  There is a specific place on the chromosome chain that must be filled with a gene.  You can only try to replace the gene with another allele or version of the gene or genes, but this is not always an even exchange.  While you may finally get that long desired gene allele combination, you may also get some other not so desirable genes as well. The main problem is that inbreeding and/or line breeding works well for genetic traits you can see, quantify or test for (such as fur type, sport ability, herding or hunting instinct), but it does not work well for traits that you cannot see, such as those that cause disease, or traits that require many genes to create.  Carriers for diseases usually show no outward signs, they only become known when they produce an affected puppy, and then it is too late.

All of these commonly used breeding practices, which are usually based on reducing the unwanted appearance or performance traits, add up to one major problem and that is an overall reduction of genetic diversity in the breed.  Yes, you could possibly get a breed where every dog is a carbon copy in look and performance, but you also inadvertently select against thousands of other genes and gene group variations you cannot see that are needed for health, longevity, naturally sustainable reproduction and more. One of the first signs of reduced genetic variation occurs when a breed starts to experience auto-immune diseases. The body’s immune system has an impact on many things such as the development of arthritis, the ability to fight off infection and ward off cancer. The gene group known as MHC (Major Histocompatibility Complex) plays an important role in the regulation of the body’s immune system.   These genes function most efficiently when the group is genetically diverse.  Take away their diversity and a long list of hard to spell disorders start to become serious problems in the breed, as well as lower resistance to infectious agents.

Once gene variations are gone from a breed’s currently available breeding stock, it is either impossible to bring them back or extreme measures such as outcrossing to another breed, have to be made.  The outcome of such an undertaking is not guaranteed to be successful or have acceptable results.
Traditions can be important, but if they make it impossible to breed healthy dogs, we need to change them. 
Breeders are not solely responsible for everything that happens to a breed.  All breed people, that is, owners and breeders alike, are accountable for what is happening in their breed. Every breed enthusiast needs to meet the responsibilities that being involved in a breed require.  Owning a purebred dog has responsibilities that go beyond providing a good home.

Cause and Effect
Late onset diseases that typically do not show symptoms until later in the dogs’ life (such as epilepsy, degenerative myelopathy, cataracts, etc.) result in genes being spread into the population every time the dog is bred before it shows symptoms. All offspring produced by an affected dog become sure carriers (if the disease is caused by a recessive gene or a polygenic group of genes) which can also then be bred and the chain of gene transmission goes on and on.

Diseases without conclusive tests to confirm them and lack of knowledge about inheritance create difficulties for anyone trying to identify which dogs carry the genes that make the trait and which don’t. 

Denial is also a major concern.  Everyone involved in a breed has to acknowledge there is a problem before anyone can begin to fix the problem. This can take years, and by then, the ‘fix’ is more difficult, if not impossible. 

Lack of honest and open communication among breeders and owners about health, temperament and performance issues prevents knowledge of the issues from being useful to further prevention. Fear of reprisal from other breed enthusiasts further prevents many people from sharing important information.

Even the most ethical and caring breeders can produce dogs with genetic diseases.  Some diseases are not clearly evident (mild epileptic seizures can be difficult to recognize, lameness can be attributed to many diseases or injury, etc.) and some owners do not report problems to the breeder that created their puppy.  Science has been able to help with DNA tests for some diseases, but DNA tests are not available for diseases that currently affect the Mudi.

Breeding Options
What are your choices?
1.  Keep on breeding with the same methods you have been using or were taught by other breeders.
2.  Wait for DNA tests to sort out the carriers.
3.  Pedigree analysis.

If breeding the way it has been done in these past decades was effective, then dogs in all breeds would be perfect breed examples and they would not have genetic defects and diseases commonly occurring amongst their members.
DNA tests are wonderful, but they take a very long time to develop and waiting for them can have serious consequences. If we stop breeding, the breed will most likely become extinct before the DNA tests we need are developed.  If we continue to breed without regard for health disorders, we face the risk of not having enough dogs left to continue the breed that are not carriers. You have to replace carriers with non-carriers in the breeding program or the number of carriers will increase. If DNA testing is not possible to replace carriers, then the reduction of affected individuals should become the goal. This can be achieved by setting risk priorities for every litter. Then use PA to choose less risky mates for those traits you wish to avoid, which requires accepting the risk for some lower risk traits to possibly appear instead (i.e. a high risk of epilepsy appearing is worse than a high risk of a bad bite).

PA is not perfect, but until there is a DNA test for everything, it is the best option we have available to eliminate or keep serious diseases to a minimum within the breeds’ population.

Coming to Terms 
Terms and abbreviations you need to know:
Gene = the basic unit of heredity
Allele = any one of two or more alternative forms of a gene
COI = coefficient of inbreeding (article link given below)
PA = pedigree analysis
Gen/gen = generation (as on a pedigree; typical pedigrees have at least 3 gens, 
     5 is what is required for PA; 10 gens minimum are used to calculate an 
     accurate COI in the Mudi, however the minimum number of gens can vary in 
     other breeds); 1st generation contains a dogs parents, 2nd generation has 
     the grandparents, 3rd has the great grandparents, etc.)
A/Affected = a dog that has a confirmed disease
C/Carrier = a dog that has produced affected offspring (one or more) or is an 
     offspring of an affected dog
SC/Suspect Carrier = grandparent of an affected dog

IMPORTANT:  Labeling a dog as affected, carrier or suspect carrier must be done responsibly!  Diseases that do not have a verified inheritance mode (dominant, recessive, etc.) or diseases that do not have a DNA test available for carrier verification, cannot be handled indiscriminately.  All littermates to affected dogs are not sure carriers and to label them as such is incorrect.  To presume that only one grandparent is the source of transmission to the carrier parent is also erroneous without DNA testing to prove connections.  Labeling a dog as an affected that has not been reliably confirmed to be affected is irresponsible.  If basic standards of PA are not followed, then the consistency which allows you to compare pedigrees will not be achieved.  Inaccurate PA providers and the scores they give will unfortunately be given the same credit as those that provide accurate scores.   The result of this will be unreliability, and PA will be blamed as ineffective.  Be aware of the standards the PA/COI source you use follows, if you value accuracy and consistency!   In other words, not all people that provide PA and/or COI are doing so with standardized methods and accurate data, therefore the results they provide may be unreliable.

Data Collection
Data sources which supply the information used in calculating PA scores:
·       Canine health databanks and registries (OFA, CERF, CHIC, etc.)
·       Kennel Club databases (Finnish and Swedish Kennel Clubs, etc.)
·      Voluntary submission from breeders and owners, often supplied under strict confidentiality agreement

Disadvantages of these data sources:
·       Information has to be continually searched for
·       Some health databanks do not publish failed results
·       Most kennel clubs do not release health information
·       Not all breeders keep records
·       Not all owners report health problems to the breeder of their dog
·       Not all breeders and owners openly share health information with others

Data Disadvantages and PA:
· Lack of information, it is not possible for a database to include every occurrence of a given trait due to the heavy reliance on the voluntary supply of info, therefore the risk scores provided may be understated or lower than is actually the case.
·    Parentage on pedigrees is not always accurate.
· Some breeds, such as the Mudi, allow dogs without pedigreed parents or known parentage, to be pedigreed and used in breeding. This leaves no information trail of family traits; these dogs can either be ‘new blood’ or ‘recycled castoffs’ (which are dogs that had pedigreed parents, but were not given a pedigree).
·    PA scores are subject to change as time passes and new info is received about trait occurrence, which means that PA needs to be repeated often (prior to planned breeding is usually recommended); each breed can have different recommendations.
·    Scores given by PA are not a guarantee that the trait is or is not carried by the dogs being analyzed; affected pups can still be produced.
·   PA can apply selection pressure against all relatives that are closely related to an affected or carrier, which has a negative side effect: choosing not to breed with close relatives can affect the gene pool negatively, because not allowing them to breed limits genetic diversity; however  allowing them to breed without knowing their carrier status can spread the genes further into the population (However, it is important to understand that any decision to withdraw a dog from breeding should not be made on the basis of PA results alone.  High PA scores, plus: poor health test results and/or DNA test results and/or unwanted structural or behavioral traits combined, might be a reason, but not PA scores alone because they are only an estimate of possible risk.).

How and where a PA provider chooses to keep all the data required is a personal decision. I use a computer program made specifically for this purpose.  Additionally, I have access to the software programs and knowledge needed to do the different PA methods, COI, etc.  I have been collecting data for my breed for more than 7 years.  The amount of data needed to do an accurate PA takes a lot of time and work to acquire.  The software program you choose needs to be adaptable for the needs of PA and you have to learn how to do PA from a qualified provider.

Know Your Enemy
Risk scores should be provided for the diseases and faults that are currently seen in each breed. It is up to the people involved in each breed to determine what those issues are.  Breed clubs, breeders and owners need to stay apprised of what is occurring in their breed and be aware of new problems that may arise. The sooner a possible problem is discovered, the easier it will be to deal with removing it, time is not on your side. Breeders and owners need to be encouraged to provide information to the source that provides PA for their breed.

Goals and Methods
Pedigree Analysis is an estimation of genetic risk in a family, also known as a relative risk assessment.  The main tool is the 5 generation pedigree of the dog or planned litter.  A pedigree is a diagram of a dog’s family that is used as a visual tool for documenting relationships, following and tracking appearance traits, disease presence, and other genetic characteristics commonly found among relatives.

The main goals of PA are:
·     To avoid producing dogs with serious health issues.
·  To increase or decrease the number of carriers of one or more traits in the population (can be traits for health, appearance, behavior, etc., however C.A. Sharp states that using her system for improving positive physical traits such as appearance and performance, is not effective in her breed; there are other systems that work well for this).
·  To plan litters that will have lower risk scores than the general population, this in turn should lower the carrier rate for the breed as a whole.
·    Conformation and performance breeders can use it to identify the dogs in the pedigree which have the traits they want to pass on such as appearance traits (fur, tail and head type), movement traits (such as side gait and style), sport capability, herding or hunting instinct, etc.  (Sharp states her method is not effective for this in her breed; there are other methods that can be used for this).
·    To determine bloodlines that are prominent.
· To study and determine if there was a past breeding strategy used (close inbreeding, line breeding, outcross.) (Sharp’s method is not used for this, other methods of analysis are required for this).

PA can also be used to:
·    Identify if there is a genetic condition in a family.
·   Aid in the diagnosis of a genetic illness.
·    Determine the mode of inheritance for a trait (a geneticist’s symbol pedigree works best for this if it is a simple inheritance trait, plus very detailed information is also required).
· Determine that a health issue is not an inherited disease (a geneticist’s symbol pedigree works best for this).
· Determine the possibility of an affected offspring appearing in a planned litter.
·   Help to identify which dogs are at risk for a genetic disorder.

Scoring System
The calculation to achieve the risk score should be based on sound research and knowledge. The methodology and mathematical equations need to be performed accurately by the provider according to a standard set of rules.  There are only two PA methods in use that I am currently aware of for disease traits in dogs and both appear to be capable of providing scores of relative risk for canine disease issues.  One method was developed by Dr. Jerold S. Bell, DVM and the other by C.A. Sharp and ASHGI.  There are other methods and types of PA, such as kinship and relationship coefficients, percentage of contribution (aka percentage of blood), and more.  While I can also perform these calculations, I mainly concentrate on the Sharp and Bell methods because disease prevention is my priority and particular interest.

Sharp/Bell PA method differences:
· Requirements: Sharp method has no specific requirements; Bell method requires for success and usefulness, a proven recessive mode of inheritance, confirmed diagnoses for the condition, verified pedigrees and the knowledge of all affected and carrier dogs.
·  The mode of inheritance of the disease: Sharp method covers all modes; Bell method requires the gene to be proven simple autosomal recessive or x-linked recessive.
·  Types of individuals used for calculation: Sharp method uses affecteds (has the disease), carriers (parents and offspring of affecteds) and suspect carriers which are all 4 grandparents (unless they are already designated as an affected or carrier which they would remain); Bell method uses affecteds (has the disease), carriers (parents and offspring of affecteds), non-affected full siblings of an affected, and full siblings to a carrier.
·  Risk differentiation: Sharp method does not differentiate between carrier or affected risk, one score is given for each trait; Bell method requires calculating carrier risk and affected risk separately, which gives a separate risk percentage for each.
·  The mathematical equation used to calculate risk: Sharp method scores are based on a modified percentage of ancestry calculation made with an Excel program, based on collected data of affecteds, carriers and suspect carriers from a 5 generation pedigree.  The Bell method uses a different mathematical equation than Sharp’s, and it is based on affecteds, carriers, and full siblings of affecteds and carriers from a 5 generation pedigree.
·  Risk score style: Sharp method uses a numerical and word index from 1-10 and low to high, along with an explanation of what the scores are equivalent to; Bell method uses percentages.
·   Selection pressure: Sharp method does not select against all close relatives to affecteds and carriers, close relatives are only included that have produced or become affected themselves, however all grandparents are assumed to be suspect carriers if they have no other designation as carrier or affected; Bell method selects against entire families as it selects not only against carriers (which is required), but his method also selects against possibly normal individuals by assessing risk (although it is lesser risk) to all non-affected full littermates of affecteds and carriers. In summary: Sharp highlights the risk of possibly normal grandparents and Bell by providing a risk factor to possibly normal full littermates of affecteds and carriers, highlights their possible risk. 
·   Identity: Sharp method does not provide to the PA recipient a pedigree of any type, name or carrier status of any dog involved in the score, therefore, individual kennels, lines, dogs, etc., cannot be selected against or for; Bell method can use public and private report types.

C.A. Sharp added the following note to this section when she reviewed this article:
“To be fair – Bell’s method is more accurate if you are dealing with a recessive trait.  It is generally applied only to a single disease within a breed.  What I’m doing covers a variety of traits with a variety of modes of inheritance.  It is “one size fits all” which, as with clothing, is convenient and repeatable, but not necessarily the best thing for every occasion.”

Sharp/Bell PA method similarities:
·     Both use 5 generations of parentage behind the dog/litter.
·     Both calculate each trait separately.
·  Both provide scores which are based on the given traits occurrence in the pedigree which are processed using a mathematical formula.
·    Both methods use only the closest relative with risk, that is, once a relative is found to be an affected, carrier (or suspect carrier in the Sharp method), the relatives behind that particular dog on the pedigree are no longer involved in the calculation; neither method adds up all the dogs on the pedigree that have a disease connection, they both only use the closest relatives and then go no further behind that one, they move on to the next relative down the pedigree, that is, only the most recent relative provides the objective risk category, you do not need the ancestors behind them to estimate transmitted risk.
·  Most scores given are a cumulative result of several relatives connected to the trait, this means that a high score is NOT usually the result of one or both parents being an affected or producer, but all the closest relatives that are on the pedigree being added together to make the resulting score.
· With both methods you can average the PA scores of the parents to arrive at an approximate score for the planned litter for any given trait.
·   Neither of the methods identifies carriers, just risk.
·   Both methods agree that PA can never clear a dog of being a carrier, if there is no known risk behind a dog in the pedigree, the relative risk of the dog is not zero, it means it is unknown.

A proper pedigree analysis should also include a COI and a risk score for each disease or fault that data is being kept for in the particular breed. Other information may also be included.
If no score is able to be given for a trait, a reason should be provided.  Reasons for not having a score can be that there is no known background of it occurring in the pedigree being analyzed or the dog or its’ parents were cleared by DNA testing.

Important Considerations: 
·     Scores given by the Sharp method are not a probability prediction.
·    The risk calculations will give you the relative risk of having a genetic trait in an individual dog or litter.
·  Relative risk analysis does not identify carriers, just risk, which means it gives risk factor estimation for a particular trait.
· The results provided by the Bell method calculation are statistical predictions and do not provide a certainty of outcome for any individual or litter.
·  To improve your chances of NOT producing either an affected or a carrier dog, you should choose a mate that offers the lowest risk factors.
·   Statistics and probabilities are not a guarantee and until we have a DNA test for a genetic condition there will always be some risk associated with any breeding.

Interpreting Scores the Sharp Way
I prefer and primarily use the Sharp method to do PA, but I can also use the Bell method.  I am more comfortable with the Sharp method and mainly use it because it is suitable to the traits in my breed.  The Sharp method can be used for any trait because it is not “mode of inheritance” dependent.  For this reason, I cover the Sharp method in more detail in this article.

The Sharp method does not distinguish between affected and carrier risk because without knowing the mode of inheritance, or in the case of polygenic inheritance or genetic predisposition, it is not possible to calculate those numbers individually.

According to the Sharp method, PA scores should ideally be no higher than1/4 the maximum score for each trait analyzed. The maximum score is 10 for any trait. In the best case scenario you would prefer to have scores of 1 or 2 for every trait, but in reality this may be very unlikely.  The goal to strive for is a score below the breed average for each trait.  However, according to C.A. Sharp, to have a realistic breed average for each trait, you need to have done at least 100 PA’s before you can start to see breed trait averages that have any meaning. (Depending on the PA method, the maximum score is 10-Sharp or 100-Bell.) These goals may not be achievable for some breeds.  Certain breed PA providers may be able to provide average scores for particular traits found in that breed acquired from doing many years of PA analyses.

A score over 1/4 of the maximum score is reason for concern.  The higher the score, the greater the level of concern becomes for that trait. Planned litters that have 1/2 the maximum score given for any serious disease should be avoided.

The best use of PA is to select mates that will lower the score of the higher scoring parent for an unwanted trait in the planned litter.  (Unlike with COI, you can average the PA scores of the parents to arrive at an approximate score for the planned litter for any given trait.)

Any dog that is given a score between 1/4 and 1/2 of the maximum score for a serious disease that occurs later in life (epilepsy, autoimmune disease, etc.), should be withheld from breeding until they are at least 4 years old so that it will be more likely that the dog is not an affected before being bred. 

If your breed provides breed averages for traits, and if a dog being analyzed exceeds the breed average, any litters planned with that dog should attempt to reduce the score to a score that is lower than the breed average as is possible.  C.A. Sharp notes that “In practice it is difficult to get more than a 1 point (10% of total) reduction in any single generation, especially where a disease is common and the average score is high.  As an example with epilepsy in Aussies, really low scoring dogs are very rare and what few I see are usually of obscure pedigree and not necessarily the best quality.  Low scores may be due as much to lack of information as anything else.  In practice most people have to settle for breeding high risk to moderate or even high risk to a lesser, but still high, risk mate.”  I hope the Mudi does not have to settle for this plan!

If there are multiple traits of concern with a particular dog, the higher risk traits should be prioritized based on their health impact.  Some traits have DNA tests available.   If a dog scores higher than 1/4 of the score, it is strongly recommended that the dog be tested before being bred.

COI Goals
Ideally the COI of a planned litter should be below 10%.  This may not be possible to achieve in all breeds or all litters.  If it is not possible, then the next solution is to produce a litter which will not exceed the average of the parents’ COI’s (example: sire-20% + dam-10% = 30% ÷ 2 = average for the litter is: 15%).  Please note, the actual COI for a litter cannot be calculated by averaging the COI of the parents, that is, you cannot take the 20% COI of the father and 10% COI of the mother and assume the litter will be the average of 15%.  COI calculation is a totally different mathematical process then PA calculation.

Damage Control
Besides PA, what else can be done to prevent the spread of disease causing genes, the production of affecteds, the reduction of gene pool diversity and still get a dog that is a nice, typical breed specimen, that is healthy and suited to its purpose?
·  Do not breed any dog until it is at least 3-4 years old, this will help to prevent breeding an affected dog and only breed them after they have passed all the recommended tests for health, temperament and purpose (such as herding and hunting instinct tests).
· Choose mates that can bring traits into your litter that you would like to have.
· Choose mates that correct your dog’s faults, but also compliment your dog’s good features.
·  Breeding ‘like to like’, without regard to pedigree, can also ‘fix’ traits the same as line breeding and/or inbreeding can; this is called assortative mating and it will emphasize or ‘fix’ traits just as inbreeding does, but without the harmful side effects of inbreeding.
· At least one or two pups from every litter in every kennel should be bred in order to maintain diversity, this is a tall order, but the meaning is that breeding a wider variety of dogs is key to maintaining diversity.  However, breeding other than healthy, typical, low risk specimens is not in the best interest of the breed; this recommendation needs to be cautiously applied.
· Uncharacteristic specimens should not be bred simply to maintain breed diversity.
·   By uniformly crossing all ‘lines’ within a breed, you eliminate the differences between them, which limits the diversity between individuals; maintaining healthy ‘lines’ or families within a breed, while occasionally crossing back and forth as a breeder sees fit to do, maintains diversity in a gene pool.  It is the different opinions of breeders as to what makes the ideal breed specimen, and the dogs they choose to breed with to create that ideal specimen, that maintains the diversity in a breed.   Forcing interbreeding among all breeders is counterproductive to maintaining diversity.   Breeding a wider selection of dogs from each litter produced in the breed is the best choice for maintaining diversity.  When a line develops a problem, it needs another line to go to for mate choices for carrier reduction; this is why crossing them all together is not recommended.
· The number of breeding offspring from high risk parents must be limited to avoid multiplying the defective gene in the breeding population.
·   Repeated breeding of known carriers to dogs in other ‘lines’ not thought to be carriers, in an attempt to reduce the frequency of the detrimental genes, is not an advisable method.  This only multiplies and spreads the gene further among the population.  If a dog is a known carrier or has a high risk of being a carrier through pedigree analysis, it should be retired from breeding and replaced with one or two of its offspring.  These offspring should only be used in a very limited number of carefully planned matings to dogs from low risk lines and these offspring should then be replaced with their offspring, with the goal of ‘losing’ the defective genes through successive generations.
·   If the majority of breeders plan matings with a carrier risk below the average of the breed, then the frequency of the defective gene will diminish in the population. This method has been proven to be successful in other breeds.
·   The most important method to manage complex inherited disorders (such as hip dysplasia and epilepsy) is to select for breadth of pedigree norms; that is, apparently normal individuals with apparently normal or mostly normal littermates have the greatest chance of carrying normal genes.  Normal individuals with affected littermates have a greater chance of carrying disease causing genes.  Normal parents who have a prevalence of normal littermates provide greater assurance of having normal puppies.
·    Inbreeding levels, breed wide, need to be reduced by eliminating the crossing of two closely related dogs to produce a litter, unless it is the only way to avoid producing a serious health issue or to prevent a bloodline from being lost.  COI calculation of every planned litter and breed wide monitoring is essential.   Keeping the COI in the majority of the litters produced under the breed average COI, will keep the breed average COI from rising.
·  Popular sire breeding should be eliminated in the breed and within kennels (all the bitches in the kennel are bred by the same 1 or 2 studs).
· Continual, repeated mating of the same two dogs to each other should be avoided as this also reduces genetic diversity.  Endless repetition of the same litter reduces available breeding combinations in the kennel as well as the breed.
·   Balance the number of sires and dams used in breeding throughout the breed and within every kennel.  Variation and equality of the sexes used in breeding is mandatory to maintain genetic diversity.  This means that as many males should be used in breeding as females, throughout the entire population.
·  Each litter which produced healthy, typical specimens should have at least 2 puppies used in further breeding to reduce genetic loss.   Half of the litter being bred would be ideal.  If only one puppy is bred from each litter, 50% of the genes are permanently lost.   If 2 pups are bred it is reduced to approximately 25% loss, etc.
·  Maintain a high average generation time for each litter produced.  In other words, the fewer the generations are that occur between the foundation stock and the current breeding stock, the better.   Genetic diversity is lost by breeding successively younger generations quickly (also known as turnover).
· Dogs (males and females) that have not reached maturity and/or have not produced a litter are unknown entities in terms of their quality and their quality as a producer.   Unproven dogs should always be crossed with older, proven mates for their first litter.  It is much easier to see what they produce when the qualities of the other parent are already known.
·  In the Mudi, it is especially important that dogs without known parents listed on their pedigree be mated for the first litter only with dogs that have known parents and have previously produced a litter.  If the resulting litter is not as good as other typical Mudi litters, further mating should be reconsidered.  At the very least, the litter should not be repeated and only another previously bred mate with known parents should be chosen.
·  Breed for litters/dogs that have balanced characteristics and avoid breeding for extremes in appearance or performance.   Breed to produce an overall balanced litter/dog that has health, vitality, temperament, performance ability, intelligence, structure and type in equal measures.
·  Monitor your litters for fitness levels.  Mating success, average litter size (for the Mudi it is 5-6 pups per litter, according to my database research), birth weight, newborn to 8 week old puppy viability, number of pups reaching adulthood (2 years of age), and length of life.  Look for declines in your litters and in the breed, which can indicate your line and/or the breed are experiencing inbreeding depression.
· Dogs that have less than good general health, cannot maintain reasonable condition or lack energy levels adequate for the breed, usually have inherent weaknesses that may be passed to their offspring.  These dogs should not be included in any breeding program. 
· Avoid using dogs that cannot breed and reproduce naturally or require repeated assistance with breeding or rearing their litter (inability to breed without repeated assistance, repeated use of AI, repeated C-sections, loss of maternal instinct, etc.)
· Males and females should not be bred more than once a year; two or more years between litters is preferred, that way health issues can be seen in the previous litter(s) before the dog is bred again.
· Try to prevent loss of original founder lines and increase those under represented.  Maintain as much of a balanced contribution as possible.
·  Breeding with populations located in other countries should be encouraged and supported.  The use of AI makes this possible.
·  Monitor and control the growth of the population.  Smooth, steady expansion from careful breeding that fulfills responsible owner demand is the best strategy for the breed.  Irregular growth patterns (overproduction followed by litter reduction) have a serious negative impact on a breeds genetic health and morale of those involved in it.

Take Home Summary
Some of the most important points you should keep in mind:
· PA can help to avoid or improve health, performance and appearance traits in your future litters or identify them in a puppy/dog you want to purchase.
·  PA done with any method cannot identify carriers or give a guarantee you will not have any serious health or other issue appear in your pups, they only give risk estimations for a trait.
· There are 2 methods of disease trait PA, only one is applicable to any trait.  The PA provider needs to follow the method rules, have the necessary database, and apply the correct math formulas to get consistency of comparison and accurate estimation scores.
· Breeding for anything less than all three of the equally important components (health, performance, appearance), that make a sound dog, will lead to further genetic variability loss and an increase in breed deficiencies and overall loss of fitness.
· Breeding practices that involve “personal” mate selection choices and line breeding/inbreeding will drive the loss of genetic variability forward as both are equally harmful and need to be avoided; assortative mating principles should be practiced instead.
· The additional steps in Damage Control are very important to improving the future for our breeds; these are not the usual recommendations, but ones that most people never thought about implementing in their breeding program or into a breed.

Breeding for sound dogs is not easy, but these recommendations should give you an idea of how and where to start.  If you would like more information about the PA services I provide, please go to the PA page tab on this blog.  Also, if you have any questions, please ask!

I would like to thank C.A. Sharp and ASHGI (Australian Shepherd Health and Genetics Institute) for their pioneering work in Pedigree Analysis and Dr. Jerome Bell, DVM, Clinical Associate Professor of Genetics at Tufts Cummings School of Veterinary Medicine, for his pedigree analysis work. 

I would also like to thank Dr. Péter Pongrácz, C.A. Sharp and Dr. Jerome Bell for their assistance and comments in writing this article and thanks to my reviewers/proof readers, Michelle Murvai, Betty Lessard and Sharon Burkhardt. 

I would also like to acknowledge and thank the translators: Dörte Kolkmeier (German) and Dr. Péter Pongrácz (Hungarian).

Links:
COI article on the Articles tab of this blog (coming soon)

References:
·   C.A. Sharp, various materials, seminar and personal communications
·   ASHGI resources found on their website (http://www.ashgi.org/)
·  The Ins and Outs of Pedigree Analysis, Genetic Diversity, and Genetic Disease Control, by Dr. Jerold S. Bell, D.V.M.
· Using Relative Risk Pedigree Analysis and Open Health Registries to Plan Matings, by Jerold S. Bell, DVM, Tufts Cummings School of Veterinary Medicine
·   A look at pedigree analysis and the closest common ancestor, by Jerold S Bell, DVM
·   Population Structure and Inbreeding from Pedigree Analysis of Purebred Dogs, Federico C. F. Calboli, Jeff Sampson, Neale Fretwell and David J. Balding
·   List of principles for 21st Century Dog Breeding, J. Jeffrey Bragg
·  Determining whether risk for sebaceous adenitis of Standard Poodles is associated with a specific DLA class II genotype; Principal Investigator: Niels C. Pedersen, Center for Companion Animal Health, UC Davis School of Veterinary Medicine