Mapping small-effect and linked quantitative trait loci for complex traits in. backcross or DH populations via a multi-locus GWAS methodology

Transcription

1 Mapping small-effect and linked quantitative trait loci for complex traits in backcross or DH populations via a multi-locus GWAS methodology Shi-Bo Wang 1,2, Yang-Jun Wen 2, Wen-Long Ren 2, Yuan-Li Ni 2, Jin Zhang 2, Jian-Ying Feng 2, Yuan-Ming Zhang 1 1 Statistical Genomics Lab, College of Plant Science and Technology, Huazhong Agricultural University, Wuhan , China. 2 State Key Laboratory of Crop Genetics and Germplasm Enhancement, Nanjing Agricultural University, Nanjing , China. Correspondence and requests for materials should be addressed to Y.-M.Z. ( soyzhang@mail.hzau.edu.cn)

2 Table S1 True and estimated values for 20 simulated QTL parameters in backcross in simulation experiment I using GCIM under various QTL effect models and K matrices (200 replicates) QTL Position Effect r 2 (%) Random QTL effect + K matrix from partial markers Fixed QTL effect + K matrix from partial markers Fixed QTL effect + K matrix from all the markers Power (%) Effect (MSE,MAD) Position (MSE) Power (%) Effect (MSE,MAD) Position (MSE) Power (%) Effect (MSE,MAD) Position (MSE) (0.04,0.16) 50.14(1.05) (0.07,0.17) 50.14(1.06) (0.03,0.14) 49.95(0.35) (0.10,0.18) (1.19) (0.10,0.18) (1.19) (0.04,0.15) (0.72) (0.05,0.18) (1.51) (0.06,0.19) (1.49) (0.05,0.18) (1.03) (0.05,0.17) (3.28) (0.05,0.17) (3.25) (0.04,0.16) (2.94) (0.19,0.35) (1.90) (0.22,0.37) (1.94) (0.20,0.35) (2.89) (0.48,0.47) (1.51) (0.44,0.47) (1.59) (0.58,0.52) (1.76) (0.09,0.25) (4.55) (0.10,0.25) (4.68) (0.07,0.22) (3.87) (0.10,0.25) (3.96) (0.11,0.26) (4.26) (0.08,0.24) (3.43) (0.03,0.14) (5.05) (0.03,0.14) (5.05) (0.03,0.13) (5.12) (0.10,0.25) (2.69) (0.10,0.25) (2.74) (0.09,0.23) (2.96) (0.34,0.42) (1.05) (0.37,0.43) (1.10) (0.15,0.31) (0.89) (0.15,0.31) (1.86) (0.15,0.31) (1.86) (0.12,0.27) (1.92) (0.03,0.14) (2.67) (0.03,0.15) (2.68) (0.03,0.14) (2.51) (0.03,0.14) (4.47) (0.03,0.14) (4.69) (0.03,0.14) (4.51) (0.06,0.20) (1.67) (0.07,0.21) (1.70) (0.05,0.18) (1.28) (0.59,0.68) (2.89) (0.58,0.67) (2.92) (0.54,0.65) (2.83) (1.03,0.88) (3.13) (1.09,0.90) (3.29) (0.97,0.81) (3.12) (0.10,0.26) (2.29) (0.10,0.27) (2.21) (0.10,0.26) (1.87) (0.03,0.14) (5.31) (0.03,0.14) (5.40) (0.03,0.12) (7.36) (0.03,0.13) (3.36) (0.03,0.13) (3.37) (0.03,0.13) (4.13) SD: standard deviation; MSE: mean squared error; MAD: mean absolute deviation. The same is true for the later tables.

3 Table S2 True and estimated values for 20 simulated QTL parameters in backcross in simulation experiment II using GCIM under various QTL effect models and K matrices (200 replicates) Random QTL effect + K matrix from partial markers Fixed QTL effect + K matrix from partial markers Fixed QTL effect + K matrix from all the markers QTL Position Effect r 2 (%) Power (%) Effect (MSE,MAD) Position (MSE) Power (%) Effect (MSE,MAD) Position (MSE) Power (%) Effect (MSE,MAD) Position (MSE) (0.17,0.32) 50.12(1.02) (0.19,0.32) 50.07(1.02) (0.14,0.30) 50.03(0.45) (0.15,0.29) (1.33) (0.13,0.29) (1.32) (0.12,0.28) (0.84) (0.11,0.26) (1.64) (0.13,0.28) (1.66) (0.12,0.27) (1.51) (0.12,0.27) (2.72) (0.12,0.28) (2.76) (0.12,0.27) (2.52) (0.31,0.40) (1.86) (0.27,0.40) (1.84) ( ) (2.78) (0.68,0.58) (1.64) (0.67,0.58) (1.56) (0.91,0.71) (2.05) (0.12,0.27) (3.16) (0.12,0.27) (2.69) (0.13,0.28) (3.01) (0.18,0.34) (4.16) (0.15,0.30) (3.86) (0.16,0.31) (3.54) (0.15,0.29) (4.99) (0.13,0.28) (5.26) (0.11,0.25) (4.78) (0.18,0.33) (2.81) (0.15,0.31) (2.93) (0.15,0.32) (3.20) (0.59,0.58) (1.13) (0.59,0.56) (1.16) (0.32,0.42) (1.01) (0.21,0.34) (2.33) (0.23,0.36) (2.43) (0.18,0.33) (2.49) (0.10,0.25) (3.74) (0.10,0.25) (3.77) (0.09,0.23) (3.48) (0.11,0.26) (4.10) (0.11,0.26) (4.09) (0.12,0.26) (4.24) (0.13,0.28) (1.72) (0.12,0.28) (1.73) (0.12,0.28) (1.44) (0.72,0.75) (2.65) (0.71,0.74) (2.65) (0.68,0.73) (2.91) (1.44,1.10) (2.95) (1.42,1.09) (2.95) (1.34,1.03) (2.73) (0.14,0.30) (1.93) (0.14,0.29) (1.95) (0.14,0.29) (1.59) (0.20,0.35) (5.15) (0.19,0.35) (4.90) (0.17,0.31) (5.09) (0.08,0.22) (4.51) (0.09,0.23) (4.67) (0.09,0.23) (4.62)

4 Table S3 True and estimated values for 20 simulated QTL parameters in backcross in simulation experiment III using GCIM under various QTL effect models and K matrices (200 replicates) Random QTL effect + K matrix from partial markers Fixed QTL effect + K matrix from partial markers Fixed QTL effect + K matrix from all the markers QTL Position Effect r 2 (%) Power (%) Effect (MSE,MAD) Position (MSE) Power (%) Effect (MSE,MAD) Position (MSE) Power (%) Effect (MSE,MAD) Position (MSE) (0.12,0.25) 50.10(1.14) (0.12,0.25) 50.09(1.15) (0.11,0.24) 50.04(0.61) (0.08,0.23) (1.67) (0.09,0.23) (1.66) (0.09,0.23) (1.31) (0.16,0.31) (2.76) (0.15,0.29) (2.76) (0.13,0.29) (2.45) (0.10,0.26) (4.02) (0.10,0.26) (4.05) (0.11,0.27) (3.87) (0.47,0.48) (2.34) (0.63,0.52) (2.41) (0.86,0.59) (3.48) (1.36,0.88) (1.95) (1.33,0.86) (1.95) (1.72,1.03) (2.22) (0.31,0.46) (5.13) (0.29,0.43) (5.04) (0.22,0.39) (5.00) (0.25,0.44) (4.55) (0.23,0.42) (3.52) (0.32,0.52) (2.71) (0.12,0.32) (3.57) (0.12,0.32) (3.44) (0.11,0.30) (4.55) (0.18,0.33) (3.91) (0.16,0.33) (4.25) (0.17,0.32) (4.30) (0.19,0.83) (1.60) (1.33,0.88) (1.56) (1.00,0.74) (1.43) (0.24,0.38) (3.26) (0.20,0.36) (3.28) (0.18,0.34) (3.55) (0.08,0.22) (4.10) (0.08,0.22) (4.04) (0.07,0.21) (4.40) (0.06,0.19) (5.20) (0.06,0.19) (5.28) (0.07,0.20) (6.50) (0.20,0.35) (2.72) (0.20,0.35) (2.71) (0.18,0.33) (2.61) (0.73,0.75) (3.45) (0.71,0.74) (3.44) (0.68,0.73) (3.63) (2.02,1.33) (5.26) (1.98,1.30) (4.77) (1.97,1.31) (4.73) (0.12,0.28) (3.27) (0.11,0.28) (3.16) (0.12,0.29) (3.14) (0.14,0.34) (7.64) (0.14,0.34) (6.96) (0.12,0.32) (7.64) (0.11,0.27) (5.06) (0.11,0.28) (4.60) (0.10,0.25) (5.50)

5 Table S4 Comparison of genome-wide composite interval mapping (CIM) with CIM and empirical Bayes in backcross with 400 individuals and 20 simulated main-effect QTL (200 replicates) Genome-wide CIM (GCIM) CIM Empirical Bayes QTL Position Effect r 2 (%) Power(%) Effect(SD,MSE,MAD) Position(SD,MSE) Power(%) Effect(SD,MSE,MAD) Position(SD,MSE) Power(%) Effect(SD,MSE,MAD) Position(SD,MSE) (0.18,0.03,0.14) 49.99(0.58,0.34) (0.28,0.08,0.22) 49.95(0.54,0.30) (0.17,0.03,0.14) 50(0,0) (0.19,0.04,0.15) 125.1(0.84,0.71) (0.37,0.15,0.31) (0.97,0.94) (0.19,0.04,0.15) (0.20,0.04) (0.23,0.05,0.18) (1.01,1.03) (0.57,0.64,0.70) (1.15,1.69) (0.20,0.04,0.16) 200(0,0) (0.20,0.04,0.16) (1.73,2.98) (0.63,0.44,0.45) (1.86,3.51) (0.19,0.04,0.15) (1.55,2.40) (0.42,0.17,0.33) (1.54,2.63) (0.31,7.04,2.64) (1.10,7.71) (0.29,0.08,0.24) (0.35,0.13) (0.68,0.50,0.47) (1.15,1.62) (0.29,3.41,1.82) (1.02,3.41) (0.30,0.09,0.23) (0.13,0.02) (0.24,0.08,0.23) (2.06,4.21) (0.03,0.01,0.12) 607(1.41,10) (0.21,0.06,0.20) (1.99,3.94) (0.26,0.09,0.24) (1.96,3.90) (0.04,0.02,0.12) 632.5(1.91,9) (0.23,0.07,0.22) (1.95,3.76) (0.15,0.03,0.13) (2.33,5.43) (0.16,0.17,0.38) (2.79,8.23) (0.14,0.02,0.12) (2.46,6.05) (0.29,0.08,0.23) (1.68,2.93) (0.46,0.30,0.39) (1.39,1.99) (0.25,0.06,0.20) (0.93,0.88) (0.40,0.16,0.31) (0.88,0.78) (0.45,9.30,3.03) (1.06,1.32) (0.32,0.10,0.25) 905(0,0) (0.33,0.11,0.26) (1.33,1.86) (0.94,0.91,0.49) (1.39,1.96) (0.27,0.07,0.22) 920(0,0) (0.17,0.03,0.14) (1.57,2.46) (0.24,0.08,0.23) (1.97,3.87) (0.16,0.03,0.13) (1.20,1.42) (0.17,0.03,0.14) (2.19,4.78) (0.26,0.26,0.44) (2.39,5.69) (0.17,0.03,0.13) (2.38,5.62) (0.24,0.06,0.19) (1.13,1.29) (0.25,0.41,0.59) (1.24,2.32) (0.25,0.06,0.20) (0.35,0.13) (0.43,0.53,0.65) (1.56,2.88) (0.44,0.33,0.49) (1.70,3.02) (0.55,0.84,0.74) (1.62,2.84) (0,0.01,0.10) 1340(0,25) (0.32,0.26,0.40) (0.29,0.08) (0.29,0.10,0.26) (1.37,1.93) (0.18,0.38,0.60) (1.69,3.51) (0.32,0.13,0.29) (0.94,0.89) (0.14,0.03,0.12) (2.74,7.44) (0.11,0.14,0.36) (2.66,7) (0.13,0.02,0.11) (3.00,8.94) (0.16,0.03,0.13) (2.05,4.29) (0.16,0.08,0.24) (2.25,5.06) (0.15,0.02,0.12) (2.24,5.24)

6 Table S5 Comparison of genome-wide composite interval mapping (CIM) with CIM and empirical Bayes in backcross with 400 individuals, 20 simulated QTL and polygenic background (200 replicates) QTL Position Effect r 2 (%) Genome-wide CIM (GCIM) CIM Empirical Bayes Power(%) Effect(SD,MSE,MAD) Position(SD,MSE) Power(%) Effect(SD,MSE,MAD) Position(SD,MSE) Power(%) Effect(SD,MSE,MAD) Position(SD,MSE) (0.37,0.14,0.30) 50.04(0.66,0.44) (0.40,0.16,0.32) 50.03(0.75,0.56) (0.38,0.14,0.30) 50.01(0.10,0.01) (0.36,0.13,0.28) (0.90,0.81) (0.48,0.26,0.42) (1.07,1.15) (0.34,0.11,0.27) 125(0,0) (0.34,0.12,0.27) (1.22,1.49) (0.61,0.62,0.66) (1.58,2.79) (0.33,0.11,0.26) (0.71,0.51) (0.34,0.12,0.27) (1.59,2.52) (0.66,0.57,0.53) (1.70,3.01) (0.32,0.10,0.25) (1.13,1.29) (0.49,0.24,0.39) (1.58,2.60) (0.62,7.84,2.73) (1.22,7.18) (0.40,0.16,0.32) (0.94,0.89) (0.83,0.76,0.65) (1.14,1.74) (0.38,3.65,1.87) (1.11,3.62) (0.48,0.23,0.35) (0.16,0.03) (0.32,0.12,0.27) (1.70,3.00) (0.20,0.04,0.14) (0.51,0.38) (0.28,0.09,0.24) (1.54,2.60) (0.35,0.16,0.31) (1.79,3.17) (0.23,0.05,0.16) 631.4(1.78,4.8) (0.31,0.11,0.27) (3.75,14.32) (0.27,0.12,0.26) (2.38,5.58) (0.26,0.33,0.52) (2.69,7.10) (0.27,0.13,0.28) (2.76,7.56) (0.39,0.16,0.33) (1.67,2.96) (0.84,1.00,0.62) (1.63,2.70) (0.34,0.12,0.28) (1.35,1.80) (0.59,0.34,0.44) (0.98,0.96) (0.69,9.33,3.00) (1.19,1.41) (0.44,0.20,0.33) (0.14,0.02) (0.41,0.17,0.33) (1.43,2.21) (0.90,0.87,0.54) (1.48,2.19) (0.38,0.14,0.30) 920(0.71,0.50) (0.31,0.10,0.24) (1.90,3.59) (0.42,0.25,0.36) (1.94,3.79) (0.30,0.09,0.24) (2.05,4.19) (0.30,0.11,0.26) (2.09,4.34) (0.36,0.43,0.55) (2.54,6.38) (0.29,0.10,0.25) (2.15,4.58) (0.34,0.12,0.27) (1.19,1.41) (0.36,0.45,0.58) (1.38,2.71) (0.35,0.12,0.28) (0.71,0.50) (0.45,0.68,0.73) (1.65,2.93) (0.25,0.14,0.31) (2.99,8.25) (0.46,0.44,0.60) (1.78,3.25) (0.52,1.32,1.03) (1.51,2.75) (0.17,0.02,0.12) (3.54,12.5) (0.42,0.59,0.68) (0.68,0.46) (0.38,0.15,0.30) (1.25,1.60) (0.22,0.24,0.44) (1.76,4.05) (0.35,0.15,0.31) (1.02,1.06) (0.27,0.17,0.33) (2.36,5.76) (0.23,0.41,0.59) (2.62,6.69) (0.26,0.15,0.28) (2.80,7.95) (0.24,0.08,0.22) (2.23,4.97) (0.22,0.20,0.38) (2.12,4.43) (0.24,0.08,0.22) (2.32,5.45)

7 Table S6 Comparison of genome-wide composite interval mapping (CIM) with CIM and empirical Bayes in backcross with 400 individuals, 20 QTL and 3 interactions (200 replicates) QTL Position Effect r 2 (%) Genome-wide CIM (GCIM) CIM Empirical Bayes Power(%) Effect(SD,MSE,MAD) Position(SD,MSE) Power(%) Effect(SD,MSE,MAD) Position(SD,MSE) Power(%) Effect(SD,MSE,MAD) Position(SD,MSE) (0.32,0.10,0.24) 50.04(0.78,0.61) (0.34,0.12,0.28) 49.97(0.74,0.55) (0.30,0.09,0.23) 50(0,0) (0.29,0.08,0.23) (1.14,1.29) (0.45,0.23,0.40) (1.11,1.24) (0.28,0.08,0.22) 125(0.50,0.25) (0.37,0.14,0.29) (1.56,2.42) (0.55,0.59,0.66) (1.61,2.9) (0.35,0.12,0.27) (1.33,1.77) (0.32,0.10,0.26) (2.02,4.12) (0.80,0.93,0.66) (1.90,3.88) (0.31,0.10,0.25) (2.17,4.74) (0.76,0.72,0.56) 350.7(1.65,3.21) (0.33,7.27,2.68) (1.20,6.73) (0.40,0.17,0.34) (0.91,0.82) (1.11,1.55,0.96) (1.32,1.95) (0.33,3.56,1.86) (1.11,4.04) (0.65,0.42,0.46) (0.50,0.25) (0.33,0.24,0.41) (2.15,5.2) (0.04,0.002,0.04) 610(0,0) (0.30,0.18,0.33) (2.35,7.35) (0.34,0.29,0.48) (1.91,3.5) (0.003,0.01,0.09) 630(0,0) (0.31,0.27,0.45) (1.39,1.92) (0.15,0.11,0.29) (2.32,5.41) (0.31,0.43,0.59) (2.51,6.15) (0.13,0.10,0.28) 800(2.70,7.14) (0.33,0.16,0.32) (1.85,4.11) (0.40,0.36,0.47) (1.41,2.10) (0.32,0.12,0.28) (1.60,2.56) (0.90,0.90,0.72) (1.19,1.41) (0.48,9.53,3.08) (1.24,1.53) (0.69,0.50,0.53) (0.42,0.17) (0.42,0.18,0.34) (1.73,3.36) (0.83,0.74,0.49) (1.59,2.52) (0.50,0.25,0.36) (1.42,2.02) (0.25,0.07,0.21) (2.13,4.57) (0.28,0.12,0.27) (2.06,4.20) (0.26,0.07,0.22) (2.37,5.61) (0.19,0.06,0.19) (2.64,6.94) (0.32,0.40,0.54) (2.47,6.04) (0.19,0.06,0.20) (3.19,10.10) (0.43,0.18,0.33) (1.61,2.61) (0.31,0.44,0.60) (1.33,2.20) (0.42,0.18,0.33) (1.56,2.42) (0.44,0.68,0.74) (1.81,3.65) (0.45,0.47,0.59) (2.21,5.31) (0.49,2.00,1.33) (1.99,4.06) (0.51,1.75,1.23) (1.63,2.88) (0.34,0.12,0.30) (1.70,3.14) (0.23,0.23,0.46) (1.75,3.47) (0.34,0.14,0.31) (1.85,3.55) (0.14,0.12,0.31) (3.06,9.42) (0.32,0.43,0.58) (3.23,10.00) (0.15,0.15,0.35) (2.86,8.13) (0.20,0.10,0.25) (2.54,6.38) (0.40,0.37,0.46) (2.85,7.93) (0.17,0.09,0.24) (2.82,8.21)

8 Table S7 Main-effect QTL identified by GCIM and some common QTL identified by CIM, ICIM and empirical Bayes. Trait Chr Posi (cm) Genome-wide CIM (GCIM) CIM Empirical Bayes (multi-marker analysis) Würschum et al. (2014) Marker interval LOD Additive r 2 (%) Marker interval LOD Additive r 2 (%) Population marker LOD Additive r 2 (%) Marker Effect r 2 (%) DS1 1A 79.8 wpt wpt A 48.9 wpt-7026 wpt A 55.8 rpt ~wpt wpt-6393~wpt3114, wpt ~ ~ ~28.78 DH07,EAW74 2A 62.1 wpt-3114~wpt DH06 wpt wpt A 17.1 tpt tpt tpt A 41.1 wpt wpt-5857~wpt ~ ~ ~13.41 DH07,EAW74 wpt A wpt A 34.5 tpt wpt wpt A 4.7 wpt-8443~wpt A 14.7 wpt A 62.4 tpt wpt0902,tpt ~ ~ ~10.89 EAW74,EAW7 8 tpt tpt A 53.2 wpt EAW78 wpt A 65.1 wpt-8377~wpt wpt EAW78 7A wpt-4489~wpt0494p7a wpt B 26.2 wpt-2526~wpt4532p1b wpt wpt

9 2B tpt tpt-1663,wpt-0100~ wpt ~ ~ ~6.00 EAW74,EAW7 8 tpt tpt B 98.7 wpt tpt B 54.9 wpt EAW74 wpt B 39.9 wpt wpt EAW78 wpt wpt B 5 wpt wpt B 53.2 wpt DH06 wpt B 62.7 wpt-5037~wpt wpt-8554~tpt ~ ~ ~20.63 DH06,DH07 6B 76.5 wpt DH07 wpt B 68.8 wpt wpt EAW74 wpt wpt R 9.9 rpt wpt DH06 rpt rpt R 44 rpt tpt , rpt ~ ~ ~7.16 EAW74,EAW7 8 rpt rpt R 66.3 rpt tpt ,rpt ~ ~ ~12.72 DH06,DH07 rpt rpt R 18.9 rpt rpt EAW78 rpt rpt R 36.5 rpt rpt ~rpt EAW74 rpt R 62.9 rpt R 81.3 rpt R 39.6 tpt rpt EAW74 tpt R 41.2 tpt rpt EAW78 rpt

10 6R 62.8 rpt EAW78 rpt R 72.7 rpt rpt R 40.7 rpt R 43.3 rpt rpt rpt DS2 2A 62.1 wpt wpt EAW74 wpt wpt A 15.6 wpt wpt A 58.2 wpt tpt ~tpt ~ ~ ~19.88 DH06,DH07,E AW78, tpt wpt A 66 wpt wpt DH06 wpt B tpt tpt ~wpt2106, wpt-2106,tpt ~ ~ ~21.52 DH06,EAW74, EAW78 2B wpt-0100~wpt rpt EAW78 tpt B 0 wpt wpt-9496~wpt-2426p3b EAW74 wpt B 98.7 wpt wpt B 39.9 wpt wpt-4936,wpt-2607~ wpt-1548,wpt ~ ~ ~27.52 DH06,EAW74, EAW78 wpt wpt B 59.1 wpt wpt-9300,wpt ~ ~ ~21.78 DH06,EAW74 wpt B 76.5 wpt tpt ~tpt-9132, wpt ~ ~ ~33.63 DH06,DH07 wpt B wpt-0406~rpt tpt DH06 rpt B 67.4 wpt wpt-9133,wpt-8312, wpt ~ ~ ~20.48 DH06,EAW74, EAW78 wpt

11 3R 35.4 rpt R 63.9 rpt tpt-4576,rpt ~ ~ ~15.30 DH07,EAW74, EAW78 rpt ,r Pt ~-0 0.1~ R 18.9 rpt rpt EAW78 rpt rpt R 41.2 rpt EAW74 rpt R 46.9 wpt-1598~rpt rpt DH07 rpt DS3 2A wpt-6393~wpt A 14.7 wpt wpt-8826,wpt-3114~ wpt ~7.96 4A 40.1 wpt wpt-5857~wpt DH07 DH06,EAW74, EAW78 wpt wpt A 2.5 wpt wpt-5787,wpt ~ ~ ~8.07 DH06,EAW74 wpt A wpt-7201~wpt wpt A 14.5 wpt A 38.8 wpt A 58.2 wpt wpt-0902,tpt ~ ~ ~13.85 DH06,EAW78 wpt wpt A 12 tpt rpt ,rpt ~ ~ ~13.84 DH06,EAW74 rpt A wpt-8377~wpt wpt B wpt-0097~wpt wpt DH06 2B wpt-6199~wpt-9958p2 B wpt-9958p2b wpt B 98.7 tpt EAW74 wpt

12 6B wpt-5408~wpt wpt B 76.5 wpt wpt DH07 wpt wpt B 68.8 wpt wpt-9798~wpt EAW74 wpt R 35.2 rpt EAW78 rpt R 65.4 rpt rpt DH07 rpt rpt R 18.9 rpt rpt EAW78 rpt rpt R 35.2 rpt rpt R 46.2 rpt rpt DH06 rpt R 40.4 rpt EAW78 rpt

13 Table S8 Probabilities in the paired t test for average statistical power and mean absolute deviation (MAD) among genome-wide composite interval mapping (CIM), CIM and empirical Bayes Case Genome-wide CIM (GCIM) & CIM Empirical Bayes & CIM Genome-wide CIM & empirical Bayes Power MAD Power MAD Power MAD Monte Carlo simulation experiment I (20 main-effect QTL were simulated) 20 main-effect QTL 2.09E E Small effect QTL E-4 Closely linked QTL Monte Carlo simulation experiment I (20 main-effect QTL along with polygenic background were simulated) 20 main-effect QTL 1.27E E Small effect QTL Closely linked QTL Monte Carlo simulation experiment I (20 main-effect QTL along with epistatic background were simulated) 20 main-effect QTL Small effect QTL Closely linked QTL

14 Table S9 QTL mapping derived from inclusive composite interval mapping (ICIM) in the first to third Monte Carlo simulation experiments (200 replicates) QTL The first simulation experiment The second simulation experiment The third simulation experiment Power (%) Effect (SD,MSE,MAD) Position (SD,MSE) Power (%) Effect (SD,MSE,MAD) Position (SD,MSE) Power (%) Effect (SD,MSE,MAD) Position (SD,MSE) (0.17,0.03,0.14) 50.02(0.26,0.07) (0.36,0.13,0.29) 50.00(0.33,0.11) (0.31,0.10,0.24) 50.02(0.50,0.25) (0.21,0.04,0.16) (0.54,0.29) (0.33,0.11,0.27) (0.62,0.39) (0.32,0.10,0.25) (0.98,0.95) (0.22,0.05,0.17) (0.74,0.54) (0.33,0.11,0.26) (1.22,1.48) (0.38,0.15,0.30) (1.37,1.90) (0.21,0.05,0.16) (1.26,1.58) (0.32,0.11,0.25) (1.14,1.30) (0.32,0.11,0.26) (1.91,3.67) (0.30,0.09,0.24) (0.74,0.55) (0.42,0.18,0.35) (1.06,1.11) (0.73,0.58,0.46) (1.23,1.51) (0.33,0.11,0.25) (0.51,0.26) (0.55,0.31,0.39) (0.62,0.38) (0.77,0.65,0.57) (1.07,1.17) (0.24,0.08,0.24) (1.37,1.83) (0.35,0.13,0.30) (1.39,1.96) (0.44,0.23,0.34) 609.9(2.51,5.7) (0.29,0.09,0.26) (1.69,2.82) (0.38,0.17,0.34) (5.45,29.66) (0.35,0.20,0.39) (0.92,1.13) (0.14,0.02,0.13) (2.20,4.79) (0.28,0.13,0.27) (2.12,4.43) (0.21,0.13,0.29) (2.35,5.43) (0.25,0.06,0.20) (1.10,1.23) (0.34,0.12,0.28) (1.31,1.72) (0.36,0.15,0.31) (1.53,2.33) (0.32,0.11,0.25) (0.34,0.12) (0.49,0.24,0.37) (0.43,0.19) (0.73,0.62,0.58) (0.74,0.54) (0.26,0.07,0.21) (0.63,0.40) (0.38,0.14,0.30) (0.98,0.96) (0.44,0.19,0.36) (1.34,1.79) (0.17,0.03,0.14) (1.28,1.63) (0.30,0.09,0.24) (2.05,4.19) (0.31,0.10,0.24) (1.88,3.51) (0.19,0.03,0.15) (1.91,3.64) (0.30,0.11,0.26) (1.96,3.81) (0.26,0.10,0.23) (2.52,6.29) (0.24,0.06,0.18) (0.81,0.66) (0.35,0.12,0.28) (1.05,1.10) (0.45,0.20,0.35) (1.31,1.71) (0.45,0.34,0.50) (1.31,2.01) (0.46,0.47,0.58) (1.43,2.32) (0.42,0.49,0.61) (1.76,3.28) (0.36,0.33,0.44) (1.50,2.23) (0.50,0.81,0.75) (1.92,3.90) (0.48,1.70,1.22) (2.21,4.67) (0.30,0.11,0.27) (1.23,1.60) (0.38,0.16,0.32) (1.35,1.92) (0.33,0.17,0.36) (1.60,3.01) (0.21,0.05,0.15) (3.02,9.07) (0.28,0.15,0.28) (2.20,4.81) (0.17,0.13,0.33) (2.71,7.20) (0.16,0.03,0.13) (1.82,3.34) (0.26,0.09,0.23) (2.11,4.42) (0.23,0.11,0.25) (2.44,6.01) QTL parameters in the three Monte Carlo simulation experiments were shown in Tables S1, S2 and S3, respectively.

15 Simulated Datasets Simulated_datasets.zip is a *.zip file that includes simulated_datasets.csv. In the simulated experiment I, the phenotypic values are in the first column and the genotypic values for all the 481 markers are in the fourth to 484th columns. In the genotypic datasets in backcross population, 1 is AA marker genotype and -1 is Aa marker genotype. In the simulated experiment II, the phenotypic values are in the second column and the genotypic values for all the 481 markers are also in the fourth to 484th columns. In the simulated experiment III, the phenotypic values are in the third column and the genotypic values for all the 481 markers are also in the fourth to 484th columns. In other words, the genotypic datasets in the second and third simulation experiments were same as those in the first simulation experiment. In the simulated_datasets.csv, each of 1 to rows indicates one individual, each of 4 to 484 columns represents one marker and 400 individuals consist of one sample. The ith sample consists of the [( i 1) 400+1]th row to (400 i )th row ( i =1,, 200 ).