*4.3. DNA Extraction and Marker Analysis*

DNA was isolated from 2 cm long leaves using the cetyl trimethylammonium bromide (CTAB) method [49]. According to the sequence variation between NIP and ZS97 (available online: http: //ricevarmap.ncpgr.cn/v2/), single nucleotide polymorphism (SNP) markers and insertion/deletion (Indel) markers in the target region were developed (Table S1, Supplementary Materials). The primers used for nucleotide variation analysis were designed according to the Nipponbare reference genome by Primer 3 (available online: http://redb.ncpgr.cn/modules/redbtools/primer3.php). Polymerase chain reaction (PCR) amplification and gel electrophoresis for marker genotype analysis were conducted following the methods described previously [50]. PCR products were sequenced by Sangon Biotech (Shanghai, China) and the sequences were analyzed using Sequencher 5.0 (Gene Codes Corporation, Ann Arbor, MI, USA).

#### *4.4. QTL Analysis and Linkage Mapping*

Germination percentage (x) such as Gmax and G3d was transformed by arcsine(x)0.5 to make the trait mean independent from the variance.

Based on the SNP genotypes, a bin was defined by a unique overlapping substitution segment from the CSSLs and used as a marker for QTL analysis. A linear ridge regression in the R package "ridge" (available online: http://www.r-project.org/) was applied for QTL analysis in the CSSL population [33]. In addition, *p* < 0.01 was set as the significance level for the presence of a putative QTL. The most significant bin was selected if several adjacent bins showed significant *p*-values. The phenotypic variance explained by each QTL (bin) was calculated using lmg in the R package named "relaimpo".

A linkage map was constructed for the CSSL-derived F2 population with an additional 10 markers on the target region. Linkage analysis and QTL validation for seed dormancy were performed using the ICIMAPPING software (version 4.1, Chinese Academy of Agriculture Sciences, Beijing, China). The presence of a QTL was declared when an LOD score was larger than 3. The additive effect and the phenotypic variation explained by each QTL were estimated by ICIMAPPING.
