Genetic Diversity and Breeding
The center of genetic and phenotypic diversity of both wild and domesticated chia populations includes the semi-temperate and temperate highlands of western Mexico and eastward spanning the trans-volcanic belt to Puebla, generally between 1,400 and 2,200 m.a.s.l. (Cahill 2004; Herna´ndez-G'omez and Miranda-Col´ın 2008; Miranda-Colin1978). Chia cultivation traditionally occurred in this region, extending into Guatemala, and in a separate area in Honduras and Nicaragua.
Limited chia germplasm is available in gene banks, and the chia that has been collected mostly comes from domesticated rather than wild populations. Two studies have evaluated genetic diversity of chia germplasm using morphological traits and genetic markers. Cahill (2001) compared diverse germplasm sources of chia (both wild and domesticated) using 33 morphological traits as well as RAPD genetic markers. The sources of the accessions used by Cahill included collections from Howard S. Gentry, housed in the agricultural institute in Chapingo, Mexico, and lines he collected as part of his doctoral thesis work. He studied 26 quantitative and seven qualitative traits ranging from leaf and floral measurements to seed color and plant growth parameters. He found relatively little morphological differences among the accessions under study, with the exception of a few major traits related to domestication. He grouped the chia accessions into four categories: wild, cultivated, primitive domesticated, and advanced domesticated. In a second study Cahill (2004) compared genetic diversity among his collection of 38 domesticated and wild chia accessions using Random Amplified Polymorphic DNA (RAPD) markers; levels of RAPD diversity within each group, HG, showed that the highest genetic diversity was found, as expected, in wild accessions (HG ¼ 0.15), with less diversity in all domesticated accessions (HG ¼ 0.10), and a subcategory was identified within the domesticated group, the modern, commercial varieties, exhibiting the least diversity remaining (HG ¼ 0.02). Of the 55 polymorphic RAPD bands observed in this study, 14 were unique to wild accessions, eight bands were found only in domesticated accessions, and one band was unique to advanced domesticated accessions. These results indicate a typical pattern of a narrowing of the genetic base of the species following domestication, with a greatly diminished amount of potential genetic variation existing in current chia lines being grown in modern production areas. Cahill (2003) found a set of traits that changed during the domestication process of chia. These include closed calyces, increased plant branching, decreased pubescence, increased inflorescence length, determinacy in flowering, more pigmentation in stems, and increased plant height. Cahill also found that seed size was larger in cultivated chia than in wild germplasm and was a part of a gigantism syndrome in advanced domesticated chia lines, as well as allelopathy. In a study of inheritance of three qualitative traits in chia, Cahill and Provance (2002) found that genes for open calyces, charcoal seed color, and pigmented stems are dominant over genes for closed calyces, white seed color, and nonpigmented stems.
In another study of genetic diversity in chia, Herna´ndez-G'omez et al. (2008) investigated variation among 22 chia germplasm sources for 23 morphological characters. Multivariate analysis of his data revealed that the chia accessions he used formed six groups, mostly related to different geographical origin. Distinct groups included chia from northern Mexico, Guerrero, Puebla, Central America (El Salvador, Guatemala and Honduras), Oaxaca, and Acatic. This study found differences between wild and cultivated chia for 19 of the 23 characters studied.
Chia has not been the subject of many modern plant breeding efforts. Improved cultivars or populations have been developed primarily by selecting lines from mixed germplasm sources, usually landraces. In general, the domesticated chia variety “Pinta” now dominates cultivation (Cahill 2005), with a few other domesticated populations or selections being grown; for example, Sahi Alba 911, Sahi Alba 912, and Sahi Alba 914 are three white seed lines developed by mass selection. Omega-3 Chia, Inc. has plant variety protection for a variety (“Omega3”) developed in Florida, obtained through mass selection from a mixed population/ landrace, likely “Pinta.”
Salvia hispanica L. is a self-pollinating plant, generally setting seed at high frequency in the absence of insects (greenhouse or mesh-covered inflorescences). Herna´ndez-G'omez et al.(2008) found much higher levels of outcrossing (over 22 %) in cultivated chia, than in wild chia (<2 %) in field studies in Mexico. Cahill (2004) reported a much lower outcrossing rate of 0.24 % in his field studies in California using a wild and domesticated line. In Kentucky over three growing seasons, outcrossing in white-flowered chia plants surrounded by blue-flowered
chia lines was in the 3–8 % range. Many insects are attracted to chia flowers, and little outcrossing has been observed under greenhouse conditions, so it is likely that entomophily is responsible for transferring pollen rather than wind. Some South American chia growers report better chia crop yields when chia is grown in areas with healthy bee populations.
Cahill and Ehdaie (2005) investigated the inheritance of seed mass in chia. They reported a 16 % increase in seed mass following one cycle of selection, but chia seed mass between wild and domesticated chia lines does not differ as much as that of other oilseed species in Lamiaceae (particularly Perilla frutescens Britt.).
Making crosses in chia is hampered by small flower size and their fragility. Many attempts at emasculation of chia flowers result in rapid floret abscission or poor success. Pollen is shed within a few hours of sunrise under greenhouse conditions. Early morning pollen transfer from plants used as males to spikes of designated female plants over several days to 2 weeks has resulted in successful crosses, with ~10 % of seed from a spike used in crosses being non-selfed (crossed) seed (personal observations). Making chia crosses is considerably easier when dominant phenotypic markers are available. Cahill and Provance (2002) used stem striation/pigmentation as a dominant marker. Breeding efforts at the University of Kentucky have used flower color as a dominant marker. By using whiteflowered plants as females in crosses with blueor purple-flowered plants as males, hybrid plants can be distinguished from self-pollinated young seedlings. Blue flowering plants produce pigmented hypocotyls in the early seedling stage of growth in bright light (1–3 weeks after germination), while white-flowering selfpollinated seedlings from the attempted cross exhibit only light green hypocotyls. Alternatively, seed from attempted crosses can be grown to flowering stage, and blue-flowering plants would be actual hybrids, while white-flowering plants would be selfed (maternal parent) progeny.
Traits of interest in a traditional chia breeding program include seed yield, flowering date, rate of maturity, lodging and shattering resistance, and disease resistance. Hybrid cultivar production could increase chia vigor and seed yield, but would likely require the use of male sterility. The use of molecular markers and other genomics tools may be productive once more genetic information is generated for chia.
With a narrow genetic base in available germplasm, other means of finding genetic variation can be used, such as mutation breeding. Possible use of male sterility for production of hybrid chia as well as novel oil/chemical profiles and plant branching pattern and height are being investigated in populations treated with gamma radiation and chemical mutagens. Additional collections of chia germplasm from its center of origin are needed and should be available to chia breeders.
At the University of Kentucky, in field plots and in greenhouse-grown chia from numerous commercial sources, variation in leaf size and shape, stem pubescence, branching pattern, spike length, seed size, flower color, and seed color have been observed, but very little variation in photoperiod response has been found. Through mutation breeding, a number of new chia lines with a range of responses to day length have been developed (Jamboonsri et al. 2012). Most of these lines are induced to flower under day lengths between 13 and 16 h and a few flower under constant illumination (i.e., they are day-length insensitive). Additional breeding efforts have been focused on studying the inheritance of traits such as photoperiod response, seed color, lodging resistance, and shattering resistance. Selection for improved yield is difficult until seed shattering and lodging problems are addressed.
