You are here

Ug99: a wheat fungus that is threatening the world's food supply

Virtually all wheat breeding during the past 100 years has employed single-gene (i.e., vertical) resistances which usually provide a complete protection, but which are liable to break down to new strains of wheat parasites. One of the worst diseases of wheat is caused by a fungus called ‘stem rust’ (Puccinia graminis) and much of the world’s wheat is protected against it by a resistance gene called Sr24.

In 1999, a new 'strain' (i.e., vertical pathotype) of the stem rust fungus was identified in Uganda and it is now known as Ug99. This strain is particularly dangerous because it can match Sr24 and, as a consequence, much of the world’s wheat is in grave danger. During the past few years, this rust spread to Kenya, Ethiopia, and the Yemen. It has now been recorded in Iran, and it is thought to have reached Pakistan. This is bad news indeed because it is only a matter of time before it reaches the Punjab and the ‘bread basket’ of Asia.

There is an urgent need for horizontal resistance breeding against wheat parasites but none of the professional wheat breeders seem prepared to tackle this. This reluctance to test anything new, and to stick with old concepts and techniques, is known as ‘scientific fundamentalism’, and it is much more common than most people realise. One of the more important objectives of the Open Plant Breeding Foundation is to promote horizontal resistance and to demonstrate just how easy horizontal resistance is to work with, and just how effective it can be.

Wheat is not the easiest crop to breed and it is not normally recommended for breeding clubs made up of amateurs. But it is a suitable project for university breeding clubs which are backed up by the university resources and expertise.

We beg agricultural colleges and universities in the countries where Ug99 is already present to establish wheat breeding clubs working with horizontal resistance to all the locally important wheat parasites. This may well prove to be the only means of overcoming the scientific fundamentalism within wheat breeding.


I visited Waklyn's Agroforestry in April of 2007 The centre is run by Prof. Martin Wolffe who developed the idea of cultivating mixtures of different cultivars to reduce infections. With some success, particularly in the old East Germany. They are now working on evolving populations of wheat having crossed 20 odd varieties that have done well in Europe during the last 20 years. During prof. Wolffe's talk on the work he was doing he didn't mention HR and the idea was to cultivate the population as a mixture and let it evolve rather than make selections. However The populations could be valuble for a HR breeding programme. As we were talking about wheat, our lecturer mentioned the breakdown of resistance in Uganda and prof. Wolffe said he had just written an article on it for the Elm Farm newsletter. When I finaly tracked it down, even he was suggesting other VR genes could be used.

I have known Martin Wolffe for many years and am well familiar with his work on mixtures which is close to the concept of multilines. However, even a mixture of VRs can break down to a complex vertical pathotype that matches every VR in the mixture. Some people (but not Martin, I think) believe that HR is no more than a mixture of VR genes, but this idea is false. This is clear once you understand that VR can only control allo-infection, and that auto-infection can be controlled only by HR. If this sounds too complicated, don't worry about it. Just bear in mind that OPBF is not recommending mixtures. Full details can be found in Self-Organising Agro-Ecosystems at the free ebooks off-site resource.

In fairness to Martin we were not degree students, he only had a short space of time in which to explain his work and was obviously most keen to get accross the ideas of growing mixtures as this was his main interest. He also works with the Sarpo potatoes, so i'm sure he's familar with HR. This is why I was so supprized by the suggestion in his article. One aspect of Prof. Wolffe's work that I found out later and is quite interesting. Is that they have introduced an element of pollen sterility (I don't know how) to increase the frequency of cross pollination. Multilines could, perhaps, die a death of a thousand cuts too, as one resistance is matched after another. If I have understood your books correctly, genetic flexability is an important part of how VRs function in wild ecosystems.

I have found a back copy of The Organic Way (No189Autumn 2007) it has an article describing the work of Dr Vandana Shiva at Navdanya Organic farm and International College for Sustanable Living. They have a seed bank with 31 local varieties of wheat, and are located in Northern India. Perhaps they would be interested in breeding resistant wheat.

Having just re-read my last post I realise I haven't made my point very well. I wasn't trying to endorse VR mixtures. The introduction of some pollen sterility to a mixture to produce new combinations still leaves it vunerable to the new pathotype that matches all the genes as you say. What's interesting for me is that the idea of a geneticly diverse population of wheat with an increased rate of cross-pollenation designed to respond to selection presure during cultivation, is very similar to continuous mass selection. The one very significant difference being the nature of the resistance. Now why can't this be used with the one pathotype tecnique to produce HR? It might not be the best method to breed HR wheat but it's a simple one. Expert help would only be need to set the thing up. Even Prof. Wolffe subcontracted out the initial cross-pollenation to the John Innes Centre. It could then be left to the farmers with only a small amount of follow up work by a co-ordinater. If the farmers gave seed to their friends as is their custom it could result in an enormous sreening population of the sort that subsistance farmers have worked with for generations.

There is another problem with VR mixtures as well. Say a new pathotype which matches all the genes emerges in one speices of parasite, and it causes a large reduction in yield. If the mixture continued to be cultivated it might develop HR to that parasite. However if there are Vertical resistances functioning to other parasites these could also be matched by a new pathotype which would lead to that VR breaking down and the HR to the other parasite being worthless as it doesn't contribute to resistance to the next parasite. So I quite understand why you don't recomend VR mixtures. But an inbreeding crop that with a bit of help from a scientist can respond to selection pressure during cultivation, is surely an oppertunity for HR breeding as well.

Your interpretation is entirely correct. Martin Beek pioneered this approach in Brazil. He had one metre-wide strips of wheat as the famale parents, and two metre-wide strips as the male parents. He used a male gametocide to make the female parents male-sterile, leaving the male parents untreated . He used a negative screening on all the most diseased plants just before anthesis. This produced a random polycross of the least susceptible plants, and a random mass selection. He also used single seed descent in hydroponics to obtain a late selection (i.e., ear to row, or family selection). He also inoculated the screening population with the most important local parasites using the one-pathotype technique as necessary. But this can get tricky if you have to have designated pathotypes for several different parasites.

Beek, M.A. (1988) Selection Procedures for Durable Resistance in Wheat, Agricultural University Wageningen Papers 88-2, 114 pp.

Waklyns is certified organic by the Soil Assocciation. So Martin isn't using male gamitocides, he wouldn't be allowed. My purpose in continuing this subject is to try and establish if this appproch could be used as an easier method of breeding wheat for amateur breeders. If the stem rust is really devastating, would it have to be inoculated? If the one pathotype tecnique was used to set the mixtures up and then left to these locally existing strains to inoculate, the progress would be slow against less important parasites due to the number of escapes and faster to more virolent P+D. Would this less involved approach be better than nothing? and do you have any contacts. at say the FAO that could organise it, given the seriousness of the problem?

Breeding for HR to Ug99 can only be undertaken in countries where it already occurs. At present, this means Uganda, Kenya, Ethiopia, Yemen, and Iran. As the rust spreads, Pakistan and India will come on board also. They would not need to use the one-pathotype technique because Ug99 would probably match everything they had.

I was not suggesting Martin should use male gametocides; if he wanted to, he could easily do so on rented land that was not certified organic. I only mentioned MGs as a useful cross-pollination technique in an inbreeding crop such as wheat.

The international scene is well aware of the dangers of Ug99, from CIMMYT down. Unfortunately, they all seem to think exclusively in terms of VR genes.

Keep on asking questions; they are one of the main reasons for this site.

Hi everyone -

I'm new to the website, and really appreciate what you're trying to do here...however, I think this article is a *touch* biased and does not fully recognize that plant breeders need to use all available tools in breeding for pathogen resistance. This includes BOTH qualitative resistance (VR) and quantitative resistance (HR). The reason many breeding programs search heavily for VR is because it is easier to integrate a few genes into existing germplasm that is adapted to local farmers' fields than it is to collectively select for dozens of them -- not because they don't want to have HR. An example given to me was that of a bowl of M&Ms. If you randomly take ten out (like the random offspring you get from crossing two plants) it is fairly easy to get a handful with a green M&M. It also might be pretty easy to end up with a green and a blue. But if you tried to specify EXACTLY what ten colors you would get beforehand, your odds are very slim. Now if you think of the different colored M&Ms as the genes that contribute to disease resistance, it becomes more clear that single genes contributing to VR are easier to randomly attain than the many genes contributing to HR. It is even more unlikely that you will find that ten gene combo if you have limited funding and need to send your populations to Ug99 screening nurseries which, as was pointed out earlier, are only in already afflicted countries.

Rust has always been a threat to wheat and VR genes have helped neutralize that threat for a long time. Plant breeders know the problems associated with this type of resistance and clearly want something better, but must work to help stop the negative impact of Ug99 using the most rapid methods available. As such it seems confrontational to make statements like: ''There is an urgent need for horizontal resistance breeding against wheat parasites but none of the professional wheat breeders seem prepared to tackle this. This reluctance to test anything new, and to stick with old concepts and techniques, is known as ‘scientific fundamentalism’, and it is much more common than most people realise.'' I think this website is a great place for people interested in breeding to learn and start their own projects, but please do not dismiss the efforts of the people working on Ug99 as though they are ignoring things.

It was also pointed out that it is difficult to cross wheat, at least in quantity. This is another problem with using a technique like recurrent selection to improve populations of self-pollinated crops. However, there are special outcrossing populations that exist in order to use recurrent selection, even in wheat. Details are given in, 'Application of a Dominant Male-Sterile Allele to the Improvement of Self-Pollinated Crops' by Sorrells & Fritz, Crop Science 1982, 22 p1033-1035. These populations have been used to select several traits for over 20 generations, and are very easy to work with. The paper is a great example of how the preconceived notion of 'selfer' and 'outcrosser' can be tossed out if you find the right germplasm to work with. (or develop it!)

best regards

My apologies for the delay in replying.

Ease of breeding is certainly a consideration but the real issue is the durability of the resistance. VR is liable to fail while HR is durable. The MM analogy is good but it refers to only one generation of recurrent mass selection. Consider having ten shots at getting green MMs. After each shot you get to keep the greens, and you will end up with all greens. The analogy begins to fall down here but the whole point about RMS is that impossible odds at the beginning become increasingly more likely with each new generation. May I suggest that you read Return to Resistance (parts 1 & 2) as this is a bit difficult to explain in one paragraph.

Your second point has validity but the fact remains that wheat breeders have consistently resisted the idea of HR. Potato breeders are more amenable but even they have done little breeding with HR. In fact, virtually all breeding for HR to blight stopped when metalaxyl came onto the market and it has only recently started up again. In the 130 years that Colorado beetle has been such a pest, only one man has attempted to breed for resistance to it. This was simply because no one could find a "good source" of resistance (i.e., a single gene, VR). This man is Peter Fisher in USA and he used RMS and population breeding. So my remark about "scientific fundamentalism" may be harsh but it is largely true.

Ethrel is quite an efficient male gametocide on wheat (see Return to Resistance for details).

Let's all of us be absolutely frank but extremely honest in our discussions. Many of these issues are both controversial and abstruse and our purpose must be to reach a good understanding of them.

With all good wishes

As I understand, Ug99 is not just a problem for wheat. Other crops too. I am abt to plant some 20 varieties ea of wheat, barley, and oats. Many are old varieties, some are untried conventional crosses with wild"ish" or completely wild species. In any event, grains are a fairly untried part of agriculture here in the Northern part of the Maritime Pacific Northwest. I expect to do some crosses in the years ahead.

1) Does it make sense to do crosses that might exhibit HR here and send them to the infected areas to trial?

2) I sort of see how we might deal with Ug99 in the case of wheat, but am I correct that there are fewer sources of resistance HR or VR for barley or oats?