Allelic quantification of STRs, where the presence of three or more alleles represents mixtures, provides a novel method to identify mixtures from unknown biological sources. The allelic stutters resulting in slightly different repeat containing products during fragment amplification can be mistaken for true alleles complicating a simple approach to mixture analysis. An algorithm based on the array of estimated stutters from known samples was developed and tuned to maximize the identification of true non-mixtures through the analysis of three pentanucleotide STRs. Laboratory simulated scenarios of needle sharing generated 58 mixture and 38 non-mixture samples that were blinded for determining the number of alleles. Through developing and applying an algorithm that additively estimates stuttering around the two highest peaks, mixtures and non-mixtures were characterized with sensitivity of 77.5, 82.7 and 58% while maintaining the high specificity of 100, 97.4 and 100 for the W, X, and Z STRs individually. When all three STRs were used collectively, the resulting sensitivity and specificity was 91.4 and 97.4%, respectively. The newly validated approach of using multiple STRs as highly informative biomarkers in unknown sample mixture analyses has potential applications in genetics, forensic science, and epidemiological studies.