Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Session 1: interaction at the production stage

Chairman A. Venkataraman
Rapporteur V. Subramanian

The improvement of nutritional quality by genetic means

Hugh Doggett

International Development Research Centre, Peradeniya. Sri Lanka



Factors associated with nutritional inhibitors, such as polyphenols in sorghum, or nutritional assets, such as carotene in maize, and which can be seen by inspection, i.e. are evident, can be dealt with effectively by plant breeding.

Inhibitors and toxic factors that can only be identified by analyses or tests, i.e. cryptic factors, require laborious procedures, but levels can often be much improved. The breeding approach should always be tried. Examples are lathyrin in grass-pea or lupinin in lupine.

Improvements in cereal protein quantity and quality also require laborious procedures. In cereals such as rice and wheat, in which increased protein levels are also associated with improved lysine levels, worthwhile progress can be made by plant breeding.

Cereals such as maize and sorghum in which increased protein levels are associated with a higher prolmine content of low quality have not been responsive to plant breeding and do not warrant expensive programmes. A low-level recurrent-selection-in-populations approach ought to be tried.


Plant breeding supplies the methodology for improving crop plants. It has been likened to the process of building, although in fact it is more like replacing the coloured marble pieces in an existing mosaic by pieces of another colour, ultimately changing the pattern substantially.

In general, plant characters can only be changed one at a time, so an order of priority is required. The second requirement is that the character under improvement should be readily identifiable by the breeder. Crop improvement is a "numbers game," in which the plant with the desired combination of characteristics may be quite rare. The breeder may have to sort through a large number of plants in order to identify the very few possessing the character he needs. If such a character is evident, a careful look at each plant will suffice to choose those desired. However. few quality characters can be seen, and they must therefore be identified by other means.

Sometimes one is lucky and finds a close linkage or high correlation with an evident character: for example, kernel weight or kernel volume may be highly correlated with oil content. The breeder picks out the plants with larger grains, which greatly reduces the numbers he must handle. The oil content of these is then found by analysis, and the plants having the highest values are selected, and grown out in the next season in head-rows, preferably replicated. A sample from the whole row is taken after harvest, and analysed. There is more grain to work with, and a better analysis can usually be made of the average oil content in the row. Growing the head-row in two replications, both of which would be analysed, helps to indicate the influence of soil conditions on oil content, and permits a better judgement on whether the high oil values selected are probably due to genetic rather than environmental effects.

Often with nutritional characters, differences are not visible and there are no good correlations with evident plant characters. Then the whole screening process requires an analytical technique, with a methodology that can handle small quantities of grain (grain is available only from single plants, and most is required for seed purposes) in large numbers. Generally, nutritional factors are difficult for the breeder to work with, both because an analytical procedure is needed, and because rapid analytical techniques for small quantities of grain are seldom available. Modern technology is steadily removing the latter difficulty.


The breeder's two priorities are yield and evident quality. Both are complex. Yield involves a series of positive genetically controlled factors that enhance grain production, which need to be accompanied by a series of resistances to the yield reducers, such as pests, diseases and drought, as well as a general tolerance of difficult soil and weather conditions. Evident quality includes grain type, hardness. colour, size, texture, and flavour. All are likely to be controlled by several to many small genes each, and so require selection in large populations. Even evident quality characters are themselves quite difficult to handle. Consumer panels for tasting and cooking evaluations are needed. Further, any breeding for nutritional improvement must not result in an adverse effect on the yield, or on the evident quality, unless it brings with it an obvious advantage, a trade-off, to the farmer. He is not going to accept lower yields unless a premium price is paid for the nutritionally superior type: nor will he want to change the evident food quality if he uses it in his own home. He would only grow it for selling.

Clearly, breeding for nutritional quality is fraught with difficulties. The plant breeder must work with a (chemical) analytical technology: the new cultivar developed must be at least as good in yield and at least as acceptable in grain appearance and good preparation as existing commercial types, or else the improved food value must be sufficient to justify premium payments: the "improved" cultivar has to be economically slightly more rewarding to the grower than are existing types. This approach to nutritional improvement has therefore not been widely used for nutritional defects, which can be easily corrected by additions to the diet.

Generally, it has been important in three areas: (a) evident characters associated with nutritional improvement; (b) inhibitors and toxic factors; and (c) protein quality and quantity.


Evident Characters

In the tropical crops, colour is often important. Sometimes it may be associated with tannin-like substances that help to confer some level of disease or pest resistance. Black beans are an illustration of this. Most people prefer the white ones, once farming standards no longer require the benefits that the colour confers.

Sorghum provides a good example of evident nutritional quality. Sorghum grains that have a persistent sub-coat, the testa, are high in polyphenols. Often the grains are brown in colour, and as the intensity of this pigment increases, so does the polyphenol content. Not only are the polyphenols bitter in taste, they also have adverse nutritional effects. Jambunathan and Mertz (1973) showed in rat-feeding trials that a low polyphenol sorghum gave an average gain per rat of 34.8 9 in 28 days, whereas a high polyphenol grain showed an actual average weight loss of 0.6 9 per rat over the same period. Oswalt (1973) demonstrated that sorghums with dark-coloured testes had a lower IVDMD than those lacking the dark testa. The mechanism of this nutritional inhibition is well known to this audience, as indeed it is to the African farmers who use these sorghums, and who have devised various methods, often involving soaking or germinating in wood-ash or lime, of reducing the tannin content of the grain in food preparation.

The plant-breeding task of selecting for white grain types is simple. and is also well known to African farmers. However, the birds eat the white types, so the people have the choice (where the birds are numerous) of no sorghum grain, or a bitter grain which the birds will not eat and which needs processing before the people can eat it themselves. The processes of beer-making greatly reduce tannin contents, and African beers are often important in the diet.

I used to work in the lake region of the United Republic of Tanzania. The small weaver birds, quelea, nest in the central part of the country, and their numbers reach many millions when the young fly. All the sorghum cultivars grown on the east side of

Shinyanga Region, adjoining the Wembere steppe where the birds nest, have dark grains with a high polyphenol content. Those in the western Shinyanga Region, away from the usual flight path and feeding grounds of the quelea, have sweet, corneous, white, high-quality grains. These are parts of the same district; most of the people living there are Wasukuma, and there is free interchange between the east and the west. Somebody has to get rid of the birds before the plant breeders can usefully get rid of the polyphenols.

There is one place where plant breeders can help: some of the high-tannin sorghums are very bitter and unpalatable to birds when the grain is in the milk stage, but the tannin level is much reduced by the time the grain has ripened. Provided that birds are not too numerous at harvest time, this grain reduces food-processing problems. The cultivar "Serene," released from Serere in Uganda and quite widely grown in the brown-grain areas of East Africa, is a normal high-tannin sorghum. A line from the 1965 cross 5D x 135 at Serere has now completed its field testing. Although only equal in performance to Serena for grain yield, the grain has this character of the tannin content falling markedly, as judged by taste. during ripening of the seed. It is now being multiplied and released from Serere, and it will be interesting to see to what extent it replaces "Serene."

Quinua (Chenopodium quinoa) is an important grain crop in South America, especially in the Altiplano of Bolivia and Peru. The situation there is exactly parallel to that of sorghum. The grains often contain saponins. and a breeding programme was started to get rid of them. However, it was apparent to me, when I visited our project there eighteen months ago, that saponins are present in grains in areas where the birds are numerous, and that good-quality grains without saponins are grown in other areas. (The saponins are removed by soaking bagged grain in the rivers.)

Another example of grain colour is, of course, yellow maize, in which the carotene can be a useful source of vitamin A. Not everyone likes a yellow maize, and other sources of vitamin A can easily be found in a mixed diet. The carotene levels at present known in sorghum do not warrant a serious breeding effort to include carotene in the breeding programme.


Inhibitors and Toxic Factors

Another class of nutritional problems are posed by toxic substances, and these are very important. The haemolytic anaemia favism from eating Vicia faba beans is mentioned in every textbook. Another important one in the subtropics and tropics is lathyrism. Lathyrus, the grass-pea, is a most useful crop plant. It grows in rough conditions better than any other pulse I know. The farmer can just broadcast it into his rice field among the stubble immediately after the rice harvest: there is no land preparation, no weeding. It comes up. grows, and gives a yield under almost any condition. It can be grazed or cut, and still regenerates to give a yield of pulse. I do not know of any legume that can do a better job for the poor farmer than Lathyrus, even half as well. It is true that the lathyrin can be destroyed by soaking and boiling the grain: but when there is pressure on the food supply, some of it gets ground up for use in bread, and then the neurotoxin begins to have its horrible effect. It is essential to get rid of the lathyrin content by breeding, and much progress has already been made here in India in achieving this, the cultivar PUSA 24 being a low-lathyrin line (Laxman Singh 1975). In Bangladesh, a pulse project supported by the International Development Research Council of Canada has taken up work on Lathyrus. An excellent range of germplasm has already been collected, including some unusually productive types, and a research worker has been here in Hyderabad for training in the techniques of lathyrin assessment.

Again, in the Altiplano of South America, lupine are a beautiful source of oil - the fields in flower are a wonderful sight - but the presence of the alkaloid lupinin requires long periods of treatment in water to prepare it as a cooking oil, and work on breeding low-lupinin lines is in hand.

Cyanogenic glycosides in sorghum furnish an example of the improvement of nutritional quality by breeding. Dhurrin is not found in ungerminated sorghum seeds, but sprouted sorghum liberates it after two days of germination, and it may be present in substantial quantities in young shoots one week after emergence from the soil. Young side branches, young tillers, and droughted and freshly rationed plants are a dangerous source of poisoning to cattle. Much progress has been made in the United States in breeding for forage sorghums and the Sudan grasses are of very low dhurrin content. Rancher was regarded as a particularly good variety (Franzke 1945), but Piper is today the most widely used of the low-HCN cultivars (Hein 1957).


Protein Quality and Quantity

The main target for nutritional improvement by breeding has been the protein quantity and quality in cereals. This was given an impetus by the concern with protein deficiency in the developing world, before it was recognized that the basic cause of protein deficiency in most situations is a food deficiency where people are not getting enough food to eat.

One attempt to develop a cereal with a high protein content that may prove to be successful is triticale, from the cross between wheat and rye. There was much fantasy in the early days, when remarkable protein contents and qualities were quoted. Little attention was paid to the fact that low grain yields are usually associated with higher protein contents, while the amino-acid profile of grains with poorly developed endosperm is often of good quality. However, now that productive types with plump grains are coming into the field, it does look as though protein content may be some 1 or 2 per cent more than in wheat, with protein quality at least as good (Hulse and Laing 1974). Triticale outyields wheat on acid soils in the tropics and subtropics in the highlands of East Africa and in South America.

The prospects of protein quality and quantity improvement in wheat itself are promising. There is a good range of variation in both protein content and lysine content. On a dry weight basis, in a large sample. mean lysine values increased from 0.33 to 0.53 per cent as the grain protein increased from 10 to 20 per cent. There was a strong positive relationship between protein content and lysine (Mattern et al. 1975).

Rice grain protein has a good amico-acid profile, with a high glutelin fraction. Milled grain usually contains some 7 per cent protein. Semi-dwarf breeding lines have been developed at the International Rice Research Institute (IRRI) which give yields comparable to those of JR-8, but with 2 per cent more protein. The higher-protein rice is nutritionally superior because of a higher level of all essential aminoacids, including lysine, in the milled rice (Juliano and Beachell, 1975).

Among the best known quality breeding programmes is that using the Opaque-2 gene in maize. A considerable amount of work has led thus far to the conclusion that Opaque-2 maize derivatives have grains that are too soft to keep well in many areas, while yields are around 10 per cent below those of the normal type. The former problem has been partly solved by the development of types with vitreous kernels, while the latter defect probably cannot be solved with the Opaque-2 gene, as the kernels of the high-lysine types are suffering from arrested development. Normal maize has around 2.5-2.8 9 lysine per 100 9 protein, while the Opaque-2 types may have around 4.0 9. It is difficult to increase total protein content usefully above 10 per cent, as the additional protein consists mainly of zein, of low nutritive value. We may conclude that farmers would do better to grow normal endosperm types, the richer farmer purchasing supplementary food rich in lysine, the poorer farmer sowing a couple of climbing been seeds at the foot of each maize plant, as he has done from time immemorial.

Sorghum is a crop greatly in need of improved protein quality. Hulse et al. (1980) state:

Recognizing the inferior nutritional quality of normal sorghum protein, the inhibition of protein by polyphenols, together with nitrogen losses that occur during domestic and small industrial processing, and probably, in storage, it is recommended that research be continued to stabilize a higher than average lysine in combination with an average (c 10% N x 5.7) protein content.

One cannot make bricks without straw: we have not yet found the straw.

In sorghum grain, the increased protein resulting from better nutrition of the plant is largely prolamine, of low nutritive value. Protein levels are low under conditions in which nitrogen is limiting at grain filling. Riley (1980) has made a thorough study of the protein situation is sorghum: the progress made in trying to use two high-lysine genes from Ethiopia, and the mutant in P-721 from Purdue. There were few grounds for encouragement. Plump seeds were unobtainable with the Ethiopian source in crosses: lysine content transferred better from P721, but yields were poor. Some figures from Riley's paper are shown in table 1.

TABLE 1. Protein in Sorghum

Nitrogen applied Line Protein %
(N x 6.25)
(% of protein)
20 kg N/ha CSH- 1 4.95 3.02
p-721 9.87 2.44
Q.50662a 9.30 2.60
170 kg N/ha CSH - 1 13.18 1.71
p-721 15.81 2.62
Q.50662a 12.49 2.48

a. Q.50662 was a "bulgy" line derived from the Ethiopian line IS 11758.
Source: Riley 1980.

The big change in lysine content with protein content of CSH-1 will be noted: the bulgy line Q.50662 had lysine levels and a response to N very similar to those of p-721. Many analyses and much effort went into this work, and there was little at the end of it showing any practical promise.

The next approach to this problem should consist of making up a population from lines with good yields drawn from diverse sources that have higher-than-average lysine contents and around 10 per cent protein when grown on land of moderate fertility. Mass selection should then be practised for ten generations, using the Udy analytical method for protein content, and progress evaluated.


Franzke, C.J.1945. Rancher, a Low Hydrocyanic Forage Sorghum. Agron. Dept. Circ. 57, South Dakota Coll. Agric. Exp. Sta. Agron. Dept.

Hein, M.A. 1957. "Sudan Grass." USDA Farmers Bulletin, 1126.

High Quality Protein Maize. 1975. Dowden, Hutchinson and Ross Inc., Stroudsburg, Pa., pp. 27-280.

Hulse, J.H, and E.M. Laing. 1974. Nutritive Value of Triticale Protein. IDRC, Ottawa.

Hulse, J.H., E.M. Laing, and O.K. Pearson 1980. Sorghum and the Millets Their Composition and Nutritive Value. Academic Press, London.

Jambunathan, R., and E.T. Mertz. 1973. "Relationship Between Tannin Levels. Rat Growth, and Distribution of Protein in Sorghum." J. Agric. Food Chem., 21, p. 692.

Juliano, B.O., and H.M. Beachell. 1975. "Status of Rice Protein Improvement." In High Quality Protein Maize. p. 457. Dowden, Hutchinson and Ross Inc., Stroudsburg, Pa

Laxman Singh. 1975. "Lathyrus (Khesari/Teoda) Cultivation in Madhya Pradesh." Tech. Bull, 26. JNKW, Jabalpur.

Mattern, P.J., V.A. Johnson, J.E. Stroike, J.W. Schmidt, L. Klepper, and R.L. Ulmer. 1975. "Status of Protein Quality improvement in Wheat." In High Quality Protein Maize. p. 387. Dowden, Hutchinson and Ross. Inc., Stroudsburg, Pa.

Oswalt, D.L. 1973. "Nutritional Quality of Sorghum Bicolor (L.) Moench Grain in Inheritance and Improvement of Protein Quality and Content in Sorghum." Res. Prog. Pep.. No. 10, p. 144. Dept. Agron. Agric. Exp. Sta. Purdue, Lafayette, Ind.

Riley, K.W. 1980. "Inheritance of Lysine Content, and Environmental Responses of High and Normal Lysine Lines of Sorghum bicolor (L) Moench in the Semi-arid Tropics of India. " Ph. D. thesis, University of Manitoba, Canada.

Contents - Previous - Next