Most of marginal lands in Indonesia are in the form of acid dry land with low available P and high Al concentrations. Development of tolerant rice varieties to P deficiency and Al toxicity is one way to increase rice production. This study aimed to select BC3F1-Pup1+Alt genotypes from three crosses based on foreground and background markers. This research was conducted at the Indonesian Center for Agricultural Biotechnology and Genetic Resources Research and Development, Bogor, from August to December 2015. The materials used were 300 genotypes of BC3F1 Dodokan-Pup1+Alt, BC3F1 Situ Bagendit-Pup1+Alt, BC3F1 Batur-Pup1+Alt, and the parents. The research included selection in modified Yoshida’s nutrient solutions (0.5 ppm P dan 60 ppm Al) followed by foreground selection and background selection. Selection using Yoshida’s nutrient solution resulted in 150 genotypes with longer root than the recipient parent in each of the BC3F1 populations. Selection with foreground markers using markers RM1361 and RM12031 produced 85 genotypes of BC3F1 Dodokan-Pup1+Alt (56.6%), 105 genotypes of BC3F1 Situ Bagendit-Pup1+Alt (70%), and 77 genotypes of BC3F1 Batur-Pup1+Alt (51.33%). Selection using background markers revealed that genotype number 116 (BC3F1 Dodokan-Pup1+Alt), number 2 (BC3F1 Situ Bagendit-Pup1+Alt), and number 129 (BC3F1 Batur-Pup1+Alt) were the best genotypes with percentage of parent recovery of 95%, 90%, and 90.5%, respectively. These three genotypes were verified to have Alt loci and had the largest genetic proportion of restoring parents. Keywords: Alt, background markers, foreground markers, Pup1, upland rice

<p>The challenges in upland rice cultivation are phosphorous (P) deficiency and aluminum (Al) toxicity, each controlled by Pup1 and Alt loci, respectively. Pyramiding the two genes into Indonesian rice varieties were previously done through Marker-Assisted Backcrossing method to obtain BC3F2 populations. The aims of this study were to analyze the BC3F2 upland rice lines containing the Alt and Pup1 loci molecularly (foreground and background analyses) and to test their phenotypic performances. Genetic materials tested included Dupa variety (donor of Alt) and three improved Indonesian genotypes (Dodokan-Pup1, Situ Bagendit-Pup1, and Batur-Pup1) as recurrent parents, Kasalath (donor of Pup1), and 300 BC3F2 lines from Dodokan-Pup1+Alt, Situ Bagendit-Pup1+Alt, and Batur-Pup1+Alt, respectively. The rice genotypes were selected individually using modified Yoshida nutrient solution, followed by foreground and background analyses. 150 out of 300 seedlings were selected and maintained until harvest in the greenhouse. Foreground analysis using markers (RM1361, RM12031, and Kas46-2) and tiller number performances resulted in 18 plants from BC3F2 Dodokan-Pup1+Alt, 30 plants from BC3F2 Situ Bagendit-Pup1+Alt, and 25 plants from BC3F2 Batur-Pup1+Alt still carrying Alt and Pup1 loci. Background analysis using molecular markers showed that the best individual lines of BC3F2 were number 56 for BC3F2 Dodokan-Pup1+Alt, number 32 or 70 for BC3F2 Situ Bagendit-Pup1+Alt, and number 20 for BC3F2 Batur-Pup1+Alt. The selected lines having both both Alt and Pup1 loci in homozygote condition with highest number of tiller per plant which are useful genetic materials for developing upland rice variety tolerance to low P and Al toxicity.</p>