Fig. 4

Additive and epistatic effects of QTL for related traits in the SSSLs and DSSLs. a Diagram of the locations of substitution segments and QTL in S1–S6 and D1 and D2. b–f Parameters of QTL in different SSSLs and DSSLs, including grain length (b), degree of chalkiness (c), panicle length (d), grain width (e), 1000-grain weight (f), and ratio of grain length to width (g). Different lowercase letters indicate a significant difference (P < 0.05), as determined by Duncan’s multiple comparison. μ: the average value of each line; ai: additive effect for each QTL controlling the trait, whose positive value shows allele from substitution segment increasing phenotypic value, while negative value decreasing one. I: epistatic effect between QTL. P < 0.05in SSSL indicates that a QTL existed in the substitution segment of the SSSL, as determined by one-way ANOVA and LSD multiple comparison with Xihui 18; P < 0.05 in DSSL indicates that an epistatic effect of Q1 × Q2 existed in DSSL, as detected by two-way ANOVA. S1 (Chr. 3 RM3417 (6.1 Mb)–RM3766 (7.4 Mb)-RM14809 (11.7 Mb)–RM7425 (13.0 Mb)); S2 (Chr. 3 RM5813 (35.2 Mb)–RM3346 (37.6 Mb)–RM1221 (40.1 Mb)); S3 (Chr. 5 RM2010 (1.2 Mb)–RM405 (3.3 Mb)–RM5874 (3.7 Mb)); S4 (Chr. 5 RM2010 (1.2 Mb)–RM405( 3.3 Mb)-RM5874 (3.7 Mb)-RM3322 (4.4 Mb)–RM3328 (5.6 Mb)); S5 (Chr. 5 RM405 (3.3 Mb)–RM5874 (3.7 Mb)-RM3322 (4.4 Mb)–RM3328 (5.6 Mb)); S6 (Chr. 11 RM26038 (1.3 Mb)–RM26045 (1.6 Mb)-RM1812 (2.2 Mb)-RM26114 (2.8 Mb)–RM6085 (3.0 Mb)); D1 (Chr. 3 RM3417 (6.1 Mb)–RM3766 (7.4 Mb)-RM14809 (11.7 Mb)–RM7425 (13.0 Mb), Chr.3 RM5813 (35.2 Mb)–RM3346 (37.6 Mb)–RM1221 (40.1 Mb)); D2 (Chr. 3 RM3417 (6.1 Mb)–RM3766 (7.4 Mb)-RM14809 (11.7 Mb)–RM7425 (13.0 Mb), Chr. 11 RM26038 (1.3 Mb)–RM26045 (1.6 Mb)-RM1812 (2.2 Mb)-RM26114 (2.8 Mb)–RM6085 (3.0 Mb)); the internal markers connected with hyphens indicate the substitution segment from the donor, whereas the markers at each end of the substitution segment linked with ‘–’ indicate that segment recombination might have occurred