Plant materials and breeding strategy
Xiaoyan22 was acquired from Researcher Zhang Li, Northwest A&F University, China. Xiaoyan22D with shorter plant height, which was discovered by ourself in population XY22, was a spontaneous mutant, and served as the recipient parent for improving the lodging ability of XY22 (Fig. S1). The other three donors as follows: (l) 91260 with two dominant complementary genes ML91260-1 and Ml91260-2 conferring resistance to B. graminis f. sp. Tritici was a gift from the Dr. Zhen Fang, Agricultural Research Center of Shaanxi  . (2): 92R137 with gene Yr26 conferring resistance to P. striiformis West. f.sp. tritici was got from Dr. Peidu Chen, Nanjing Agricultural University. (3) ZhengNong 16 with two tightly linked Glu-D1 genes encoding high-molecular-weight glutenin subunits (HMW-GS), Glu-D1d (Dx5+Dy10) was developed by Tiwen Lei , Academy of agriculture and forestry of Zhengzhou  .These four parents were crossed in pairs to produce three single-cross F1 hybrids, then intercrossed to produce a double-cross F1 hybrid (DCHF1). "The genotype Ml91260-1, Ml91260-2, Yr26" of DCHF1 plants was verified with applying molecular marker analysis and field evaluation of resistance." The verified DCHF1 with genotype (Ml91260-1, Ml91260-2, Yr26) was crossed with the third single-cross F1 hybrids produce a three-cross F1 hybrid (TCHF1) for pyramiding of disease resistance genes and HMW-GS Dx5+Dy10. The plants of TCHF1 with genotype (Ml91260-1, Ml91260-2, Yr26, Dx5+Dy10) were backcrossed with XY22D two times to produce plants which agronomic traits similar to those of XY22D (TCHF1BC1 to TCHF1BC2) with MAS and field disease resistance evaluation. The subsequent two generations (TCHF1BC2F1 to TCHF1BC2F2) were raised through selfing, with MAS exercised and agronomic traits investigated in each generation, eventually leading to selection of six improved groups, including eighteen lines, designated Pyramiding lines (PYLs) and Plants (PYL1-6; Plant 1-18, Fig. S1). The eighteen plants were multiplied individually to produce sufficient progenies for evaluation of phenotypic traits, yield and grain quality.
DNA extraction and PCR amplification
In each generation, marker assisted selection (MAS) was exercised using simple sequence repeat (SSR) markers for tracking specific genes. The total DNA for polymerase chain reaction (PCR) amplification of every generation of plant material was extracted according to the procedure of Sharp PJ et al . Specific SSR primers which are adjacent to ML91260-1, ML92160-2, Yr26 and Dx5+Dy10 [17,19,22]，were adapted for PCR amplification (Table 1). Total reaction mixture was 15 μL containing: 100 ng of genomic DNA, 1X Taq DNA polymerase buffer, 10 pmol of forward and reverse primers, 2.5 mM of each dNTP and 0.15 U of Taq DNA polymerase (Takara，USA). Template DNA was initially denatured at 95℃ for 5 min, prior to 32 cycles of denaturation at 94 ℃ for 1min, annealing at 50-65℃ for 45s and extension at 72℃ for 30 s. In the final step, the reaction mixture was incubated at 72 ℃ for 10 min before completion. After amplification, 3 μL of restriction buffer (10×) was added to the PCR products, which were separated on a 4 % agarose gel.
Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE )
SDS-PAGE for analysis of HMW glutenin subunits Dx5+Dy10 in the TCHF1BC2F3
was conducted as described by Damania with minor modifications, the separating gel concentration is 12% and the concentrated glue is 8% . Protein fractions was extracted from half seeds of each plant and electrophoresis was performed at a constant voltage of 200 V with a low concentration of running buffer [16.7 mM Tris 127.9 mM glycine, 0.07 % (w/v) SDS]. Electrophoresis was continued until protein bands carrying low molecular weight components reached near the end of the gel.
Evaluation of agronomic traits and disease resistance of XY22 and the PYLs
Six PYLs and control XY22 were sown at the experiment station of Northwest A&F University in ShaanXi province with three replications from 2015 to 2017．Each entry consisted of eight rows (4 m) with 161 seeds per row and a row-to-row distance of 25 cm. Seeds of XiNong2000, a variety highly susceptible to yellow rust and powdery mildew were sown at the border as a guard row. Leaves of plants in the guard row were inoculated with Pst isolate E09 and Egt isolates TiaoZhong33 at shooting period. Inoculations were performed by brushing condidia from neighboring sporulating susceptible plants of XiNong2000 from booting stage to heading stage. E09 pathotypes caused severe powdery mildew epidemics in the wheat-growing regions of north China and TiaoZhong33 pathotypes caused severe yellow rust epidemics in the wheat-growing regions of south GanSu and central ShaanXi, China . Data were recorded on the following nine traits: (1) plant height at shooting period and maturation period (cm), (2) powdery mildew resistance, (3) yellow rust resistance, (4) grain quality, (5) kernel number of 30 spikes, (6) 1000-grain weight (g), (7) grain yield (t/ha), (8) productive tillers in 1m2, (9) days to heading and flowering. The PW severity was scored using the 0-9 scale as Blanco A et al  descried. 0-4 scale represented resistance, and 5-9 scale was susceptible. The disease symptom was scored using the 0-4 scale according to resistance level as described by Mcintosh RA et al  for YR. 0-2 * scale indicated resistance, 3-4 scale was susceptible. A strong wind, resulting in the lodging of XY22, occurred in May of 2017. Thus, the data was investigated only in 2015 and 2016.
Flour Quality analysis of XY22 and PYLs
For each PYL, 1.5 kg was analyzed for flour quality. Wheat gluten content was analyzed by Glutograph-E (Germany). Wheat bulk density was measured with a bulk density meter (BLH-5000, China). Dough formation time and stability time were analyzed by a wheat extensograph（Farinograph-E, Germany).
Comparison of the means of the data of all nine agronomic traits in two years were performed using the analysis of variance. The data on grain yield (kg) was converted into grain yield (t/ha) for further statistical analyses. P values less than 0.05 were considered to be statistically significant. Statistical analysis was performed using the SPSS 19.0 software (IBM Inc. Armonk, NY, USA).