Research Article | Open Access

Volume 2025 |Article ID 100051 | https://doi.org/10.1016/j.plaphe.2025.100051

RGB imaging and computer vision-based approaches for identifying spike number loci for wheat

Lei Li,^1,2,3,9 Muhammad Adeel Hassan,^4,5,9 Duoxia Wang,¹ Guoliang Wan,¹ Sahila Beegum,^4,5 Awais Rasheed,^1,7,8 Xianchun Xia,¹ Yong He,³ Yong Zhang,^1,2 Zhonghu He,^1,9 Jindong Liu ,^1,2 and Yonggui Xiao ¹

¹State Key Laboratory of Crop Gene Resources and Breeding, Institute of Crop Sciences, National Wheat Improvement Centre, Chinese Academy of Agricultural Sciences (CAAS), Beijing, 100081, China
²Zhongyuan Research Center, Chinese Academy of Agricultural Sciences, Xinxiang, 453519, China
³Institute of Environment and Sustainable Development in Agriculture, Chinese Academy of Agricultural Sciences, Beijing, 100081, China
⁴Adaptive Cropping System Laboratory, USDA-ARS, Beltsville, MD, 20705, USA
⁵Oak Ridge Institute for Science and Education, Oak Ridge, TN, 37830, USA
⁶Nebraska Water Center, Robert B. Daugherty Water for Food Global Institute, 2021 Transformation Drive, University of Nebraska, Lincoln, NE, 68588, USA
⁷Department of Plant Sciences, Quaid-i-Azam University Islamabad, Pakistan
⁸International Maize and Wheat Improvement Centre (CIMMYT) China Office, c/o CAAS, Beijing, 100081, China
⁹Lei Li and Muhammad Adeel Hassan contributed equally to this work

Abstract

The spike number (SN) is an important trait that significantly impacts grain yield in wheat. Manual counting of SN is time-consuming, hindering large-scale breeding efforts. Hence, there is an urgent need to develop efficient and accurate methodologies for SN counting. A YOLOX algorithm was used to determine the optimal growth stage for developing wheat spike detection models among recombinant inbred lines (RILs) across Zhongmai 175 × Lunxuan 987 and a diverse panel of 166 cultivars. We subsequently increased the precision of spike identification by developing a new YOLOX-P algorithm that incorporates the convolutional block attention module and increasing the resolution of the input images. We also used these SN data to identify underlying loci in the Zhongmai 578 × Jimai 22 RIL population. The results revealed that the late grain-filling stage presented the highest precision among the SN detection models, with accuracies ranging from 91.8 to 95.02 %. The improved YOLOX-P algorithm demonstrated higher mean average precision scores (5.30−5.99 %) and F1 scores (0.06) than did the YOLOX algorithm when it was applied to the same subsets. Three new SN loci, namely, QSN.caas-4A2, QSN.caas-4D and QSN.caas-5B2, were identified using the 50k SNP arrays. Two kompetitive allele-specific PCR markers linked with QSN.caas-4A2 and QSN.caas-5B2 were developed, and their genetic effects were validated in a diverse panel of 166 cultivars. These findings provide useful tools for high-throughput identification of SNs and novel loci in wheat.