Just run a few dogs through it, very interesting.

Link please, or do we find it from main KC site easily?

http://www.the-kennel-club.org.uk/services/public/mateselect/

Have spent half the morning playing - as tohme says, very interesting :)

Yes a link would be good as I have just tried to find it on KC site and as per usual I can never find what I'm looking for on there! :-(
Paula xxx

That was easy to do.

I now have a question, the coefficient for my breed is 7.6% and it states you are to aim for lower in the pups, does it mean lower than the parents or lower than the average?

If the pups were to be higher than average then how much higher is acceptable?

None of my paper matings using my own dogs ended up being higher than average.

Seems totally weird to me -says my breed inbreeding is 2.2%. 2.2%? Maybe if you mated them to a completely different breed altogether. Where on EARTH did they get that figure from? Makes no sense at all in a breed where everyone is closely related to everyone and we can never find stud dogs that aren't too close.  I can only assume the large number of working dogs are changing the overall picture, but of course those dogs are not relevant when it comes to show lines.

Does sound very low, what scores did your own dogs get and were they more realistic?

I've just put through every single dog in my breed that was placed at Crufts this year (although I noticed some dogs cannot be found even when I know I have spelt it correctly -one being my own) and they vary from 0.8 % for the lowest (import sired) to 22.7 % for the highest, with the majority being between 3.8 and 13. Only 3 in total were 2 or lower.

I wonder how others will find the figures in their breeds?

I've been playing with this all afternoon and have found it interesting. I especially like the "Prediction of the inbreeding coefficients of puppies from a hypothetical mating" part.
Paula xxx

I am gobsmcked that the COI for the labrador is 6.4% when that for my own breed with a small gene pool it is 4.1%.

There is no advice on what they base these averages on, is it for all breeding in it's database from the year dot, or for more recent pairings, as in my breed our numbers/gene pool have reduced markedly over the last 20 years, and I would bet that the average COI has increased for litters in that time.

My own dogs came out at Elka born 1992 3.5%, her daughter Tula born 1995 0%(imported sire).  Her daughter Kizi 7%, her daughter Jozi born 1999 6.9%, her daughter Lexi born 2003 by Import sire 0%, her daughter Myka born 2006 by overseas sire 0.2%, and Inka by her half brother born 2008 13.7%.  Her daughter Safi born 2011 from my USA trip 0.4%.

I think they need to re-do the breed COI's giving more info, the overall COI, then say that for the last 5 years. 

Perhaps have a tool where COI's for different decades can be found so that the bottlenecks can be pinpointed.

I think this will be really useful, I found it through their Facebook page, which is here http://www.facebook.com/Kclovesdogs. The link they gave to mate Select was this won: http://www.mateselect.org.uk

Also, they have given out this email address for questions - might be helpful mateselect@thekennelclub.org.uk

> I can only assume the large number of working dogs are changing the overall picture, but of course those dogs are not relevant when it comes to show lines.

Then that's going to skew the figures for a lot of other breeds as well.... any where there's a working/show split. I wonder if the coefficients for breeds which don't have a working split - purely show - are higher?

based on my own dogs except where using an import/quarantine born parent or overseas stud the COI's have been 6.9% - 13.7%.

Because of the small gene pool I have of course used imports and travelled abroad so the dogs resulting from that have had COI's of 0 - 3.5% so maybe that is how they arrive at such a low average COI.

It gives an artificial picture of the state of the breeds gene pool, as we have reached the stage of having to import every other generation to keep at reasonably low inbreeding levels.

I wonder if the coefficients for breeds which don't have a working split - purely show - are higher?

Well Papillon and Cavalier were more than twice as high as Malinois (5.3 and 5.2) , with so many more dogs around so yes, I'd bet that is the case.

i tried it with my girl and some that i like i put tilly in who was 9.1% the dog is 11.0% predicted puppies would be 1.8% with breed average being 6.3% i thought this quite low can any one explain what the percentage means (the ones under dog names )

Thanks

Nessa

Edited to say This is with a black
If i put with another red dog highest was 10.4% and another is 6.4% (line-bred ) 

well im thick and dont really understand what is a good(score) higher or lower than the figure they give on the right?

Wow i just did some of our very well known dogs in our breed and most came out around 29 - 35% even though it says breed average 18%, i guess with alot of my breed theyre very old school and also a very small gene pool.

The higher the number, the more inbred the dog is. I did a test right now and put in two LITTERMATES as potential parents and that came out at 34.8 %. You could only ever get 100 % if you mated full sibling to full sibling for several generations.

My breed COI is 5.8%

My girl - 5.3% over 13 gen.
My oldest boy - 2.3% over 15 gen.
My youngest boy - 4.8% over 14 gen.

A lady I met a bit ago wanted to put her pet boy (with many faults) to my girl, I said no. Just did his name and he has 14.8% over 14 gen.

channel 5 news ran the kc mate select story tonight, Caroline Cisko told people to go to ABS members what a hoot , in my breed some of the biggest puppy farmers are ABS members , sorry slightly off topic

You have to check how many generations the COI is for.  Tollers have a breed average of 2.4% over 8 generations.

It's interesting because for border collies this type of program has been available for some time, but of course the more outcross matings the diverse the dogs will look and the more problems may be shown... interesting that there will be 'swings and roundabouts'.  eg HD will show up more in outcrosses due to the complexity of the genes causing hereditary HD.  There is a lot of discussion over using carrier dogs or not.  In Australia in border collies where they've had big problems with a storage disease (CL) and also a blood disease called TNS - both fatal.  Before DNA testing they very successfully identified most of the CL carriers and removed them from the genepool (fair enough because it's a fatal disease and it needs eliminating) using the dogs which were CL clear.  Unfortunately it was some of these dogs which carried a previously unknown condition TNS which also kills.  Now fortunately there is testing for both and these awful diseases can be avoided, although there was recently a poor pup born with TNS where the 'breeder' allegedly had no idea about the disease.  It is a cautionary tale of gene pools though.  We are fortunate in the UK that border collies don't have a high number of CL carriers, although I think we have found more TNS than expected.

My breed is 9% which is a figure that doesn't surprise me. My foundation bitch (born 1980) doesn't show up but her daughter (born 1983) does. My last bitch, out of 2 imports comes up as 1.9% and her daughter, mated to a UK dog as 0%. We don't have any working lines and, until recently, a very limited gene pool.

it didn't give generations except on individuals

Interesting - I've run through one of my dogs, KC reckon that Border Collies are 3.4%, Lexie is 1.8% and the other collie site reckons she's 6% 'inbred' so not sure what to make of it!  Her daughter is 0.7% according to KC and 5% on the other site.

It would be useful to be able to 'find' dogs as well!  You can do this on the collie site, try out matings and see outcomes for everything from hip scores to colours.

There are health tests missing too :-( Rosie's TNS test isn't on there!

It would be useful to be able to 'find' dogs as well!  You can do this on the collie site, try out matings and see outcomes for everything from hip scores to colours.

That would be a good idea, I was hoping mate select would be more like you put in your bitches name and a list of potential stud dogs comes up with details of their tests, COI and so on.

2.2%? Maybe if you mated them to a completely different breed altogether

LOL my thoughts exactly--I tested this out with my boy and every bitch in our breed that's been placed in the show ring in the last year and got results ranging from 12-30+%, or in other words all above breed average.

Having said that I noticed that a good 30% of the dogs in our breed that placed at one of the champ shows this year (with a numerically strong entry) were all sired by the same dog.

i thought mate select was going to tell you which handy stud dog down the road would suit your bitch :( :( or did they give up on that idea in the end?

>i thought mate select was going to tell you which handy stud dog down the road would suit your bitch  

As the KC doesn't keep a record of stud dogs how on earth could they do that?

> As the KC doesn't keep a record of stud dogs how on earth could they do that?

That was part of the PR and news soundbites when it first came out - it was billed as a dog dating service, don't know if this was by the KC or the may the media picked it up.  One of the lines I remember was that it would "help owners of bitches find stud dogs near to them".  So glad it is a very different kind of tool.

The reason some breeds are coming up with unexpectedly low COIs is because Mate Select treats the earliest dogs in the electronic database as unrelated founders. Same goes for imports, even if they descend from exported UK dogs.

The full worldwide Toller database shows that the average COI for the breed is around 25 per cent - (making Tollers, on average, the equivalent of full sibs to each other).  According to Mate Select, though, the breed COI is only 2.4 per cent.

Jemima

I have noticed that each different dog % varied in how many generations it went back....this gives very different results for each dog. I think it ought to go back a set number (say 8 gens) of generations to give a more realistic result.

Unfortunately is can't, certainly not in a breed like my own where we have imports almost every second generation because we have a numerically small breed.

All the results I got back told me they went back 15 or 16 or more generations, but had complete pedigrees from 4 - 6 generations.

Even my overseas trip to Finland came back with a COI 0.2% and USA one 0.4% so they are taking into account known ancestors/overseas dogs where known.

Ok then within a breed then, as one generation difference can give such a different results.

> The reason some breeds are coming up with unexpectedly low COIs is because Mate Select treats the earliest dogs in the electronic database as unrelated founders.

How far back do the dogs in the electronic database go? If what you say above is correct, then the older KC recognised breeds would have high COIs, and the newer breeds lower?

Do we know exactly how many dogs are in the database across all breeds? I'm thinking that in some of the more popular breeds it could take years to input every known pedigree... so if there is an actual point at which they haven't added every known ancestor, it's not going to give accurate results.

I think rather than just see a COI % come up with every dog's name or proposed pairing, I'd like to see how many generations and individuals were used to get to that figure, and a link to a viewable pedigree so that I could check that for myself, make sure it had been recorded correctly (typos are a fact of life!!) and if there were any blanks.

I think the electronic database started in 1980 (don't quote me as not sure right, but it's around then). It does, however, include dogs that go back further than that as it includes pedigree data for those early dogs (which reaches further back than the start of the database). The Imperial paper mentioned KC electronic records going back to around 1970 and that paper included data going up to 2006. That  was 5.7 million dogs from 207 breeds - with a median of 3443 dogs per breed. Assuming around quarter a mill reg'd dogs a year, that means the KC database now includes around 6 million dogs. A lot!

I think Mate Select is probably most useful for UK breeds with few imports - and, more generally, for making people/breeders more aware of COI.

It is important to note, however, that COI is not really a measure of genetic diversity of a breed.

Jemima

> >It is important to note, however, that COI is not really a measure of genetic diversity of a breed.<br />

Is it going to be a helpful measure of anything though, if the KCs figures aren't accurate....?

This is a rather interesting article: http://www.canine-genetics.com/pgbreed.htm

Is it going to be a helpful measure of anything though, if the KCs figures aren't accurate....?

Yes I think so.

It is limited, but the figures will be accurate enough for some breeds and they do provide a benchmark. The advice is to breed litters have a COI lower than the breed average (similar to the hip scheme, I guess) and if that is done, it will help slow the loss of genetic diversity (although not "prevent" it, as I see the KC press release claimed).

Jemima

Also worth reading from Dr John Armstrong (sadly no longer with us):

http://www.canine-genetics.com/cake.htm

It is John Armstrong who did the study that showed that the least inbred Poodles lived, on average, four years longer than the most inbred ones.

Jemima

Yes i should have provided the link to the whole group of articles: http://www.canine-genetics.com/Default.htm

The advice is to breed litters have a COI lower than the breed average (similar to the hip scheme, I guess)

That was my point earlier. To do that with my breed you'd have to crossbreed or use totally unrelated imports for every litter. Almost all matings would be above otherwise. And my breed doesn't have any health problems.

???????????????????

Tohme with the Malinois COE showing as 2.2 %, if you had to go lower than that, for the show dogs you'd have to do as above- see my earlier posts. There just aren't any dogs available to use.

I was really referring to your statement that Malinois have no health problems.......... I thought that hip scoring and eye testing were recommended for the breed?

I realise that is not what you are referring to in your post re the Mate Select test, but what about epilepsy?  Do you think that the system could help there? I was mulling this over with GSD as well, wondering if, as there is no test for epilepsy if this could be useful or is monitoring the best method on a case by case basis?