Chasing unicorns with the neXtgen Index

Chasing unicorns

I started with Ferg about a year ago. I got told to pack a suitcase and jump in my Subaru, and after a few laps around Australia I seem to have come back here with just a suitcase for myself.

So today I'm talking about chasing unicorns. Basically, in the ram industry, and from travelling around, we're always looking for the perfect ram. So far I haven't found it, Ferguson hasn't found it, no one's found it. But we're trying to make tools that can help us breed or create a ram that's more profitable for our enterprises. And the fact of the matter is that the unicorn for each person's place here is probably looking slightly different.

I was told not to ask too many questions when I'm up here, but I don't listen to instructions very well. So I wanted to see where you guys start and how you pick rams to match your breeding objectives. What are some of the things you look at? Is it the catalogue, or is it phenotype when you're selecting a ram? Don't all go at once.

Structure, yep. So sorting through the catalogue beforehand. Is anyone using an index? Don't all put your hands up at once. Not too many. Not all like Ferg, then.

Why one index can't suit every farm

Right, so using indexes. The point of indexes is that traditionally in the sheep industry we've selected on one or two traits and gotten ourselves into a lot of trouble, and ended up with an unbalanced animal. So the original point of indexes was to make balanced genetic gain and decrease that risk of single-trait selection, because obviously there are negatively correlated traits with each trait you select. Thankfully Chad covered a bit of this earlier. If you're not looking at your breech wrinkle, you're going to have a lot of other traits associated with that. It's a similar thing if you're just chasing fleece weight: there are a lot of traits correlated with that, and if you're not watching carefully those traits are going to move in another direction.

So indexes were originally created to try and balance those traits. I'll just run through some of the indexes we've got at the moment, and I think we might find why some of them haven't been so widely taken up.

This is one of the first ones. I don't know how well you can see at the back, but this is the Fibre Production Plus, and what's on here is the weightings of the index, for the top 10 per cent of animals that come through when using that index. So the top 10 per cent of animals on FP+ have 45 per cent more wool and are 37 per cent finer, with a little bit on lambs weaned and a little weighting on staple length and staple strength. So that's the weighting for the FP+. As an index, it's basically looking at two traits.

We switch over to the Merino Production indexes, the MP and the MP+. These indexes are really heavily weighted on fleece weight, and then a little bit on fibre diameter and a little bit on yearling weight. Once again, pretty dominated by one trait. Then if we move to our Dual Purpose indexes, which are probably the closest of the Sheep Genetics indexes to what we're breeding for, for a lot of our clients and here, they're using clean fleece weight, trying to hold fibre diameter, a little bit of yearling weight, and then trying to push reproduction, with only a little bit on eye muscle depth.

The issue with these indexes is that it's pretty hard to take one index and use it as a blanket and say, right, this index is going to be perfect for all environments across Australia and New Zealand. You can see just on the slide here we've got a lot of different environments. There are people here today from four states and some very different production systems. I'm sure the genetics that'll work well in Victoria, where at the moment it's about minus three and blowing a gale, are probably pretty different to what's going to work in central Queensland, where it's 35 and sunny.

As I said, what works here won't necessarily work there. So we can't just throw a blanket index that's weighting three or four traits and expect it to produce in every environment in Australia. There's more than one trait we're looking for.

One thing that affects what we're looking for in our genetics is the environment we're managing our sheep in. The other thing is actually our management system. You could compare yourself to your neighbour, or even across the hill here, and you could be running completely different genetics and management systems. So what works here may not even necessarily work right next door, as you can see with this little fella on the screen.

That's the big part about indexes. Say one person next door is trying to finish lambs at nine or 12 months of age for a wether kill, and another person is selling weaners straight off mum. Those people are going to have different genetic objectives for where they want their growth rates to sit. It's a similar thing if one person is in Victoria and he might have a high worm burden or drench resistance. He's going to have to place a larger emphasis on worm egg count than, say, someone out at Tottenham, where Dave's from. So our environment and our management systems affect what genetics we should be chasing.

Building the neXtgen Index

After degrading indexes quickly, there is a light at the end of the tunnel, because we've decided we'll develop a tool of our own. Between a couple of scientists, a few economic models and one opinionated Excel nerd, we got together and decided we'd start to create a neXtgen Index. As you saw before, the traditional indexes are basically weighting two to four traits. In this index we've weighted 17 traits, but we have the capacity to weight up to 32 traits, and in the future probably more. So we're trying to take in and weight a lot more traits to build that more balanced modern Merino that we hear about so often. If this doesn't get you excited, I don't know what will. I don't know why you're sitting back there looking at me like you've just seen a ghost.

This here is the weightings that we've got for the neXtgen Index, and you can see the 17 traits down the side. From sitting down, we decided that yearling weight, or hogget weight, for a lot of people is too late if you're trying to join ewe lambs, or if you're trying to finish wethers. You want them finishing at around nine months of age generally, or even earlier. So we put a lot more emphasis on weaning weight and post-weaning weight, and then a negative emphasis on adult weight. I should just say quickly that this is the top 10 per cent of animals that this index pulls up in the Sheep Genetics database, compared to the average.

So those animals are basically five or six per cent better for weaning weight and post-weaning weight, and then back to only four per cent higher for their adult weight. We're trying to breed those animals that have an early punch and then plateau out. We know that's got a correlation with fat and muscle, so we've put a large emphasis on fat and muscle to help cap that adult weight and breed for that genetic condition score.

We're quite happy to keep fleece weight where it is. So basically we're holding fleece weight and fibre diameter. Fibre diameter may look like it's going up slightly, but once again the industry average is minus one, and it's pretty hard to make a generic index when there are people with breeding objectives down to 16 micron or even lower, right through to 22, or even higher in some of the composites. So we tried to make one that was quite generic and sits around 18.5 to 19 micron. We're quite happy if yearling and adult fibre diameter is sitting slightly higher than minus one, so it's probably going to be sitting at minus 0.7 or 0.8.

We've got a negative emphasis on fibre diameter CV. Reproduction may look like it's weighted heavily there, but due to the traits we've weighted, it's kind of come along for the ride. We've put a slight weighting on staple strength, and there's quite a large emphasis on the breech traits, but we've put more of an emphasis on dag. If you're where Dave is, you're probably going to want more of an emphasis on wrinkle and cover. But if you're in Victoria at the moment, in winter, everything's under a pretty heavy dag, so you're going to want dag probably nearly as important as wrinkle, if not more, and you'll want a lot larger emphasis on it.

Then we've split up the three new reproduction traits. For anyone here who hasn't been following Sheep Genetics recently, NLW, which is number of lambs weaned, has recently changed to weaning rate. It's on a new analysis system, and that weaning rate is made up of conception, litter size and ewe rearing ability.

So we've split those three up, and we tried to put heavier weighting on conception and ewe rearing ability. But due to the fertility of the animals that came up in the top 10 per cent, it actually made them all quite equal in their weightings, at around five per cent. We put a little emphasis on maternal weaning weight, because we're trying to breed a maternal Merino and we don't really want that going negative. And then we've put an emphasis on fleece rot. I'm sure most people here, especially if you're in New South Wales, have experienced a pretty wet year this year. Driving up yesterday, we nearly needed a boat.

So there's been a fairly heavy emphasis on fleece rot and colour this year, and we've put a negative weighting on that. But once again, if you're out at Hay, you mightn't need that emphasis, so we can take that out. You may think there are a few traits we've left out, such as the eating quality traits, like shear force and IMF. We're trying to make this for the commercial producer at the moment, so until those premiums are there, and we've got people in Merinos sending animals in to get killed and tested for those traits, we're probably quite happy to leave them out.

That doesn't mean we're not focusing on those traits with our stud producers. For example, if we're making an index for Mumblebone, you can see they're quite focused on keeping those eating quality traits in check, so we'll weight those traits when we're working with them. This one here is just a generic index. Once again, I'm not saying this is perfect, but we think it's a good step in the right direction, and we'll probably be changing it in the future if there are premiums in that space.

Tailoring the index to your farm

Does anyone have any questions on this before I go on? Have we got time? Sure. Right, anyone got any questions about what's on the board? No one's as excited as I am. Cool.

Yes, we can. What I'll be talking about in my next slide is that we can change the weightings on each one of these traits to suit individual farmers. As I've been saying, where you guys are, in a high rainfall environment, you might want more of an emphasis on worms and dag. And if you've had heavy pressure on your wools, you might want more on fleece rot. So we can put all those weightings in. If you're finishing animals for a kill job, you might want a bit more weighting on early growth. So for individual producers we can change the weightings on that index, and for stud producers we can change it to perfectly fit their breeding objective.

Yes. So the scale of it is very similar to the other indexes. What animals come to the top is quite different, though, because we've included such a large number of traits. It's really biased towards studs that have actually gone out and recorded and measured all of these traits. That's been really good, and hopefully it's going to encourage a lot of producers, if they want to rank well in this index, to start measuring those traits. I'm happy to be argued against, but I think every trait on this board is important in the modern Merino.

So hopefully once we start rolling this index out, more people start recording more traits. Because a stud like this one has recorded a lot of traits, when we checked the rams out there last night, nearly 50 of those rams were in the top five per cent, which I think would sit quite differently to other industry indexes. Anyone else? That'll do.

Right, so what this index can do for you. As I said, we've taken this index and made it generic, so we can compare between studs on traits that are more relevant to what a lot of our clients' breeding objectives are, which is a more modern Merino. But as I was saying before, we can tailor it to your breeding objectives, so we're trying to maximise genetic gain towards your breeding objective. As Ferg was talking about before, we can get a flock profile test, and then if we set some targets for where you want your ASBVs to be, we can tweak that index to try and move your flock average towards your target.

Then if you're selecting rams based off that index, obviously taking into account structure and other phenotypic traits, you're going to maximise genetic gain to try to reach your breeding objective as fast as possible. That's where I see the major advantage in this index. Say you get a flock profile test back and you go, my sheep are from quite a traditional base. They're cutting the amount of wool I want them to cut, they're at the micron where I want them to sit, they've got plenty of growth, but I'm probably lacking in carcase traits and breech traits. We can more heavily weight those traits so you can reach that objective faster.

As I said, it's really handy to rank ram catalogues. If you're going through a ram catalogue before your sale day, we can rank those rams on this index, and then basically all you've got to do is go through, pick out the top 40 rams and go and inspect them. It's also pretty handy to search for sires. I'll spend probably a bit too much time, Ferg will tell you, sitting on this index searching through the Sheep Genetics database looking for the next unicorn.

And then also MateSel. I don't know how many people here know what MateSel is, but it's basically mating software on your computer which looks to maximise genetic gain while minimising inbreeding. Originally we had to put in an index, say one of those Sheep Genetics indexes, or a particular trait that you wanted to maximise, and MateSel would maximise the genetic gain towards that index or trait while minimising inbreeding. This tool's been really helpful in MateSel, because we can put in that exact breeding objective and then try and maximise genetic gain towards that objective while minimising inbreeding.

Finding the right ram team

So, finding the right ram or the right rams for you. I guess what you're all thinking is, oh, we're suddenly going to release this index and it's going to rank the rams from most useful to least useful, and everyone's going to go for that higher-ranked ram. In a way, that's actually probably what we're trying to do, because if an index isn't reflecting what people here want to buy, then it's probably not reflecting most people's breeding objectives. So yes, that's part of the point. If the price of the rams isn't reflecting the Sheep Genetics indexes, then I think those indexes are quite different to what people's breeding objectives are in here.

But if we can change the index for people that are sitting in different areas under different management, we're going to end up with different rams coming to the top of that index. We might have someone out in the sheep-wheat belt who's trying to finish wethers at seven or eight months of age, and he's going to want a bit more growth in his animals. But he's realised that his animals are always in good nick, because he's always got a lot of feed and he's not pushing them too hard on stocking rate. So he might be looking for some of these bigger animals.

And then if we ended up with someone in a high stocking rate area around Victoria, they might go, well, we need to really punch in fat and muscle, we need to be looking at dag, we need to be looking at worms, because they're the most important traits economically for that enterprise. They might be selecting a completely different ram team.

A bit about where we're hoping to take this index in the future. We've been doing a lot of structural scoring this year, and we're hoping to make a breeding value for that, so in the future we're hoping to include structural traits in this index. That actually takes out a bit of the work on sale day again. You can put in a cut-off and say, if the structural trait is too bad, say past a score of three, then we're not going to look at it and we won't put that in the index.

We're also hoping in the future to put things like footrot in there, for people in Victoria. These poor Victorians are getting a hammering. We can also bring Tasmania into that.

Another use for this index is that we can rank raw data for commercial enterprises. As the use of EID ramps up, we're finding people are starting to want to use that technology and record more data on their ewe flock. You saw Dave Greig putting up ewes on the screen that he basically knows by name. That's the power of EID: you can be picking those ewes that are performing every time.

So it's really handy in ranking raw data for commercial enterprises. We've been doing a bit of this, where we get, say, growth, the scanning of those ewes, and what their condition score is at each joining, and then we can create a ranking for those ewes when we're classing. All we're doing is assessing the ewe on its phenotypic merit, but we've also just got a ranking in front of us. That's why I say one of the big advantages of this index is actually in classing: combining that raw data, which you can collect pretty easily through EID, with the phenotypic look of the animal.

What the top-ranked rams look like

I didn't want to throw too many numbers up today, but I did think I'd put this up. Here I've got the average of a few traits for the sale day rams that are out in the pens. I ranked them on the index and pulled out the top 10 per cent. You can see that the top 10 per cent have about two kilos more growth, and nearly half a mil more fat and muscle. They're pretty consistent on fleece weight, they're slightly finer, they've got a lot lower CV and they're higher in staple strength. They're slightly plainer and they've got a lot less dag.

Most importantly, they've got a higher weaning rate. And then that last one on the end there is fleece rot, so they're also going to withstand the weather better. So you can see that this index is really handy in pulling out the top-performing animals.

Just to tease you even more, I grabbed the five top-ranked animals that are out there today. I don't know if you want to go and have a look at these, but I know I already have. You can see the top five animals it's pulled out, and I know I'd be pretty content bringing a trailer here and taking any of those home.

I know it's lunchtime now and I've probably held you up for long enough, but I just wanted to say good luck going out there and chasing those unicorns.

Questions from the audience

Audience member: How does the accuracy of the ASBVs fit in with the index?

Henry Hickson: At the moment, even if it's got a low accuracy, that estimated breeding value is the best estimate of where the animal sits, so we're not actually worried about the accuracies. If you're going to pick a ram team of, say, those five animals, even if the accuracies are low, if you buy five rams the chances are they're going to throw to the average. So accuracy is probably more important if you're just trying to pick one ram, but what we're hoping this index will really help with is picking a ram team.

It brings up the top animals that will help maximise profitability in your enterprise, so then you can pick a team that's got a good average EBV. Also, Sheep Genetics aren't displaying breeding values any more for animals that don't meet their threshold for accuracy, so those traits just won't be included in the index if they don't come through.