Breeding programs are often constrained by the genetic diversity of the parental lines, even though these lines can be a rare source of unique alleles not found elsewhere. Therefore, identifying these rare alleles is crucial for keeping them in the breeding programs while introducing new genetic resources. The growing amount of whole genome sequenced data has made Genome‐wide Association Study (GWAS) dominant in investigations to find causal genes for all crops, including legumes. However, GWAS often fails to predict more than one causative mutation (CM) in multiple alleles of a single causal gene. Consequently, multiple alleles complicate breeding when not recognized by a single associated marker, which typically identifies only the most frequent CM and discriminates against the others. In this work, we focus on adopting recent applied genomics methods to identify multiple independent alleles and rare alleles in soybean as a model for other legumes. We predicted, identified, and confirmed a new and extremely rare CM for the loss of black pigmentation in the soybean seed coat and hilum color R gene, the Q25fs. The deletion of eight bases leads to a frameshift, a premature stop codon, and a truncated, nonfunctional protein. Our results also suggest a possibly new gene or an allele of the seed coat color inhibitor I gene. Using a soybean model, we demonstrate how applied genomics methods can accelerate pre‐breeding, and additionally, we discuss the potential for adopting these methods for application to other legumes.