Research Center for Global Food Security and Ecosystems (GFE)

How to breed crops: Utilizing or removing their evolutionary history?

The Tambora eruption in 1815 was a catalyst for plant breeding sciences and incentive for the creation of the University of Hohenheim. What has followed since then was the invention of many high-yielding and resilient plant breeds. Today, plant breeders are utilizing the evolutionary history of a crop in their effort to breed new cultivars. However, some approaches actually try to reverse the evolutionary trajectory to make crops cooperative and increase population performance. But is this really the way to go?

When thinking about the roots of agricultural science and the incredible ways it has emerged, we always recall the story of the University of Hohenheim. The largest volcano eruption ever recorded happened in 1815 in Indonesia. The Tambora volcano released so much ash and sulfur that it caused a global climatic phenomenon that later became known as the “year without a summer.” Despite the distance, the persistent sulphate aerosol veil which dimmed the sunlight affected not only Asia and Northern America but also Southern Germany. In 1816, Germany experienced an extremely cold summer. The crop failure that ensued resulted in a massive famine. In order to make agriculture more sustainable and cataclysm-independent, new knowledge centers were needed. Thus, the University of Hohenheim, founded in 1818, became such an institution.

Even though people far away from Indonesia were hardly capable of associating the Tambora eruption with the “year without summer,” they realized that nature has an almost infinite capacity to surprise, and not always in a good sense. So, they boldly started their pursuit to outmanoeuvre nature. They improved crop fertility, bred more resilient varieties, and adapted crops to different climatic environments. For instance, the University of Hohenheim has recently introduced a chia cultivar that can be grown in the temperate climate of Germany. The scientific progress in fighting for food security and breeding crop varieties that are resilient to climate fluctuations within the last decades and centuries is astonishing. But much of it rests on the idea of creating traits uncharacteristic for naturally evolved ecosystems, which seems logical: Survival in nature requires different “abilities” of plants in comparison to maximizing crop yields in an artificially created ecosystem and for agricultural purposes.

However, what if the mechanisms of adaptation in nature are also of use for breeding resilient crops? In view of this, the lecture of Prof. Dr. Ford Denison hosted within the Rethinking Agriculture seminar series gave the audience fresh and original insights into modern science: How about not only trying to breed and select better crop varieties, but in doing so also utilize evolution as a tool?

According to Prof. Dr. Denison, understanding the evolutionary selection processes of the past may be useful for understanding a plant’s adaptive advantages and what constrains them. Nevertheless, when it comes to applied agriculture, the focus is generally on immediate issues rather than possibilities of evolutionary history. So, would it not be appropriate to ask: What if evolutionary history is a hint not only to the past but also to the future development of agricultural science?

All of the examples of plant breeding brought forward by Prof. Dr. Denison involved plant traits that would have been useful in nature but had never been selected for because the traits reduced the individual competitiveness of a plant in an environment of "selfish" conspecifics. Thus, the plant carrying this beneficial trait never made it far enough to spread it within the population. But let us assume that this plant was not outcompeted, and instead, the entire population behaved cooperatively to the benefit of all. You guessed it: It is time we talk about group selection.

The concept of group selection was created by Charles Darwin, contentiously debated in the second half of the 20th century, eventually discarded, and is now re-emerging by the new name of multilevel selection (although let us stick with the term “group selection” for simplicity’s sake). Very briefly, group selection is a type of natural selection that (supposedly) acts on all members of a given group instead of the individual. While the theory of individual selection asserts that natural selection only acts on the genetic level, group selection can act on multiple levels simultaneously, the genetic level as well as higher levels, such as the population.

The idea of group selection is probably best summarized by Charles Darwin who claimed that “[...] moral men might not do any better than immoral men but that tribes of moral men would certainly ‘have an immense advantage’ over fractious bands of pirates.” In other words, an altruistic, cooperating group of individuals has better reproductive success than a group made up of selfish, competitive individuals, even if that entails constraining oneself in one way or another. But we shall reserve debates about altruism for sociobiologists and mediocre barbecue parties and talk about group selection in an agricultural context instead.

In agriculture, group selection makes a lot of sense because “[y]ield, the most fundamental agronomic variable, is a characteristic of the population, not the individual” (Weiner et al. 2010). However, as we have seen in Prof. Dr. Denison’s talk, there is no universally applicable recipe for plant breeding aimed at population performance. An excellent example that illustrates breeding for group selection is Weiner et al.’s (2010) high density, uniform cropping system. They propose that weed suppression does not have to involve any kind of chemical intervention. Crops that grow densely to one another easily out-shade and kill unwanted agricultural weeds. In addition, the uniform sowing pattern makes sure that soil resources are evenly distributed among the population.

The problem is that densely sown plants also heavily shade one another. This triggers a plastic mechanism, the shade avoidance response, resulting in extreme elongation growth of the crop. As a consequence of this competitive behavior, the crop allocates less energy for its seeds and invests more energy into vegetative growth, thereby decreasing yield and increasing the risk of lodging. Thus, for this concept to work, the crop must behave cooperatively and abandon the race toward the light. However, such cooperative behavior would very unlikely occur in nature since a crop without the shade avoidance response cannot persist in the shade of other plants for very long. But what if the entire population lacks the shade avoidance response? In this spirit, plant breeders are currently trying to breed crops that do not show this behavior.

But is this still Darwinian agriculture, or is it not the antithesis of it? Even Weiner et al. (2010) called the approach of breeding for “[...] characteristics that maximize population, not necessarily individual, performance [...] different from natural selection.“ Apart from a few very unique examples, group selection has barely ever passed the test of natural selection. Thus, concrete examples of how a cooperative, group-selected organism interacts with species of different trophic levels and taxonomic groups are lacking. To be more specific, it is hard to tell what might happen in the long run on a population or ecosystem level if an ancient evolutionary mechanism like the shade avoidance response is removed from a plant’s gene pool, which forces the individuals of a population to cooperate with one another. In our opinion, this is not utilizing evolutionary mechanisms to improve yield anymore, it is the opposite, and it leaves a bitter taste to an otherwise impressive plant breeding effort. On the other hand, we desperately need new and creative ways of practising agriculture, so let us just see how it plays out, shall we? After all, what is the worst that could happen?

Here you may find:

Authors: Thomas Köhler, Maria Kunle


Comments:

dnpatrice2001

I recommended the use of evolutionary mechanisms for the breeding of new cactus variety that is ground tolerant and endures high yields. But would the new variety effectively allow farmers to overcome land shortage and to obtain high yields?

In the framework of my master thesis research, we analyzed cactus value chain in OUED EL MALEH valley of Casablanca Morocco. From the research field work, the majority of farmers do not respect the planting density in force in the region. Could this apparent lack of control over technical production routes be the result of the experiences acquired by farmers? Is it a strategy unconsciously adopted by farmers to exploit the apparent differences in the expression level of cactus trees in order to optimize the cactus yield? To better understand how farming practices may affect yield in the region, we run a single regression analysis by taking as a dependent variable, the yield of cactus farm and as Independent variable, the planting density (the number of cactus trees at that farm). The results highlight on the logarithmic basis a significant relationship between these variables. However, we remarked on the normal basis that there is no significant relationship between the yield of a cactus farm and the planting density. From the analyses, we concluded that all cactus trees of a farm are not responsible for the high yield. So there is a differentiation at the expression level of different cactus trees in terms of productivity and ground tolerance. Consequently, we recommended to run experiments within cactus farms to determine the cactus trees that are responsible for high yields, to carry out in-depth laboratory analysis in order to find out at the level of the ecotype, the DNA sequence that codes for high yields and ground tolerance. The identification of productivity and ground tolerant traits in the genomic sequence of cactus plants could allow farmers to produce more with the land shortage which is decreasing because of urbanisation.

After Prof. Dr. Denison's lecture, I realize that those recommendations are not consistent enough. The use of evolutionary history for breeding of new cactus variety would not guarantee high yield to farmers, because the yield of a farm depends on the population, not the individual. The problem is that densely planted cactus heavily Shade one another. This provokes a plastic mechanism (as mentionned by Prof. Dr. Denison), the Shade avoidance response, resulting in extrême elongation growth of the plant. As a consequence of this competitive behaviour, the plant allocates less energy for the formation of fruits and invests more energy into vegetative growth, thereby decreasing yield.

Today is the question: What should be done to make all cactus trees of a farm significantly contribute to high yields?

As explained by Prof. Dr. Denison, we need to reverse the evolutionary trajectory to make crops cooperative and increase population performance. Accordingly, the ancient evolutionary mechanism like the Shade avoidance response should be removed from cactus' gene pool, which forces the indivuduals of a cactus population to cooperate with one another. To this end, knowledge on how crops function from the enzyme to ecosystem scale is required for this impressive plant breeding effort.

 

Jacob Weiner

I think the very interesting and knowledgeable comments by Thomas Köhler and Maria Kunle display a slight misunderstanding of the evolutionary arguments that Ford Denison and I have been advancing. The goal of an evolutionary approach to agriculture, and especially plant breeding, is to learn from nature, not to emulate nature. Evolution via natural selection maximizes individual fitness, but this does not maximize population performance (or even survival). This give agriculturalists the opportunity to do things that are good for agricultural production, but that nature would never produce. Köhler and Kunle are correct to point out that there are potential riskes in developing palnts that would not past the test of natural selection, but this has been what plant breeding for higher yields has always been about. In a short recent paper ("Looking in the wrong direction for higher-yielding crop genotypes", 2019 ,Trends in Plant Science - available via my web pages), I point out that high-yielding crops are not "better" plants - on the contrary - they are worse, even impaired, plants, that could not even survive without the environment created by farmers. The ability to survive and reproduce in nature has been "traded-off" for the ability to produce high yields under "luxury" conditions.

Another point I would like to make here is that, despite the assumptions of some plant breeders, most of the increases in yield to date have been due to changes in management (e.g. fertilizers) which required new genotypes. The "Green revolution" is an example. Fertilizers were introduced to south Asia in the 1950s and 60s. The local crop varieties could not deal with high nitrogen levels: they grew too much and fell over. New, shorter varieties, which would not fall over, were needed. (Only later did plant breeders notice that the shorter varieties had higher yields even if the older varieties were propped up so they did not fall over.) The point is that the new genotypes would not have produced higher yields under the previous, low-nutrient conditions. Advances in agriculture have been due to genotype x management interactions, not genotypes alone.

 

denis036

I agree that what I propose is not using "evolutionary mechanisms" - natural selection tends to undermine cooperation among plants - but rather evolutionary inferences (thinking about ways humans can improve crops in ways missed by natural selection, as opposed to assuming we can design a more-efficient enzyme that natural selection has). Weiner's work on crops that cooperate to shade weeds (even at some individual cost) is a good example. Another example of something missed by natural selection (not covered in my talk) is breeding for benefits to future crops. For example, leaving persistent root channels or enriching the soil with more-beneficial microbes would not be favored by natural selection if it had any individual cost, but it's something we could perhaps breed for. I'm working on this particularly with respect to symbiotic nitrogen fixation.