Cumulative and different genetic effects contributed to yield heterosis using maternal and paternal backcross populations in Upland cotton

www.nature.com/scientificreports OPEN Received: 5 January 2018 Accepted: 20 February 2019 Published: xx xx xxxx Cumulative and different genetic effects contributed to yield heterosis using maternal and paternal backcross populations in Upland cotton Lingling Ma1, Yumei Wang2, Babar Ijaz1 & Jinping Hua1 Heterosis has been utilized in commercial production, but the heterosis mechanism has remained vague. Hybrid cotton is suitable to dissect the heterosis mechanism. In order to explore the genetic basis of heterosis in Upland cotton, we generated paternal and maternal backcross (BC/P and BC/M) populations. Data for yield and yield-component traits were collected over 2 years in three replicated BC/P field trials and four replicated BC/M field trials. At single-locus level, 26 and 27 QTLs were identified in BC/P and BC/M populations, respectively. Six QTLs shared in both BC populations. A total of 27 heterotic loci were detected. Partial dominant and over-dominant QTLs mainly determined yield heterosis in the BC/P and BC/M populations. QTLs for different traits displayed varied genetic effects in two BC populations. Eleven heterotic loci overlapped with QTLs but no common heterotic locus was detected in both BC populations. We resolved the 333 kb (48 genes) and 516 kb (25 genes) physical intervals based on 16 QTL clusters and 35 common QTLs, respectively, in more than one environment or population. We also identified 189 epistatic QTLs and a number of QTL × environment interactions in two BC populations and the corresponding MPH datasets. The results indicated that cumulative effects contributed to yield heterosis in Upland cotton, including epistasis, QTL × environment interaction, additive, partial dominance and over-dominance. Heterosis refers to the phenomenon of F1 hybrids performing better over their parents in yield, quality and adaptation. Dominance, over-dominance and epistasis hypotheses have been proposed to explain the heterosis mechanism. The three hypotheses demonstrated complementarity between dominant alleles and deleterious recessive alleles1,2, superiority of heterozygote3,4 or mimicry over-dominance with repulsion-phase linkage of favorable alleles5,6 and interactions among non-allelic genes7–9, respectively. Some previous studies reported the major role of dominance effect on heterosis in rice10 and maize11. However, over-dominance had also been detected as the primary genetic basis of heterosis for decades, such as in maize12,13, rice14, rapeseed15 and tomato16. A SFT gene was reported to cause strong yield heterosis governing by over-dominance in plant architecture17. The Dw3 gene contributed to heterosis for plant height in a way of repulsion linkage in sorghum18. Additionally, novel experimental design and molecular quantitative genetics approach has been used to elucidate the importance of epistasis at two-locus level in rice during the past decades19–22. Recently, Jiang et al. suggested that dominance effects played a less prominent role than epistatic effects in grain-yield heterosis in wheat by developing a quantitative genetic framework23. Recombinant inbred line (RIL) population is available to dissect additive and additive × additive effects but lacks heterozygous genotypes to dissect dominance and dominance-related genetic effects. So attempts have been reported by constructing testcross (TC) or backcross (BC) populations and immortalized F2 (IF2) population to 1 Laboratory of Cotton Genetics, Genomics and Breeding/Beijing Key Laboratory of Crop Genetic Improvement/Key Laboratory of Crop Heterosis and Utilization of Ministry of Education, College of Agronomy and Biotechnology, China Agricultural University, Beijing, 100193, China. 2Institute of Cash Crops, Hubei Academy of Agricultural Sciences, Wuhan, 430064, Hubei, China. Correspondence and requests for materials should be addressed to J.H. (email: ) Scientific Reports | (2019) 9:3984 | https://doi.org/10.1038/s41598-019-40611-9 1 www.nature.com/scientificreports/ www.nature.com/scientificreports create heterozygotes in rice10,14,24,25, maize11,13,26 and cotton27,28. Dominance complementation was considered as the major genetic basis of heterosis in rice because heterozygotes were superior to respective homozygotes in a BC1F1 population in rice10. Most QTLs underlying grain yield displayed apparent over-dominance effects, and little difference was observed between heterozygous genotypes of nine families of hybrids in three RIL populations in maize13. Epistasis and over-dominance were the major genetic bases of inbreeding depression and heterosis for grain and biomass yield by using five rice populations14. Heterotic effects and dominance × dominance interaction explained the genetic basis of heterosis in an IF2 population deriving from an elite rice hybrid21. Over-dominance, pseudo-over-dominance and epistasis were estimated as important contributors to yield heterosis using a high-density genetic map in rice22. Among main-effect QTLs and digenic epistatic QTLs pairs, over-dominant loci were the most important than additive, complete and partially dominant loci in two BC populations based on one same RIL population in rice24. Dominance, over-dominance and epistasis contributed to the genetic basis of heterosis using a 3,184 bin-map in an IF2 population in maize29. Moreover, new strategy of heterotic haplotype capture was proposed to trace novel heterozygous chromosome blocks for breeding30. A recent report proved that the new statistical models of QTL mapping can completely dissect large-scale time course data in post-genome era31. Yield potential has always been a vital target of plant breeding in cotton. Significant yield heterosis was previously reported in cotton27. It is also a major breeding solution to exploit heterosis for improving yield on Upland cotton. For decades, 271 QTLs were available for yield and yield-component traits in the CottonGen database32. Among 4268 QTLs in Cotton QTLdb database33, 87, 59, 98, 169 and 305 QTLs were detected for seed-cotton yield, lint yield, boll number per plant, boll weight and lint percentage, respectively. However, less QTL have been resolved for seed-cotton yield, lint yield and boll number per plant due to complex experiment management, heavy workload and highly accurate data. The qSCYchr07a displayed strong over-dominance effect and the qSCYchr07c explained 38.96% of phenotypic variation for seed-cotton yield34. A total of 14 QTLs were identified for seed-cotton yield, lint-cotton yield and lint percentage in a RIL population of Upland cotton35. Dominance and over-dominance contributed to seed-cotton yield heterosis in an IF2 population derived from a heterotic hybrid of ‘XZM 2’ in Upland cotton36. Heterotic QTL analysis suggested that over-dominance mainly contributed to cotton yield heterosis37. Twenty-three QTLs were identified for boll weight and lint percentage in an intraspecific population of Upland cotton38. Fifty-eight (...truncated)