The results returned from the 3130 sequencer were analysed using GeneMapper® ID v3.2 to determine which
samples were suitable for further use. For the one-contributor this website investigation eight replicates of each of three conditions were created (Table 2). The conditions were created to investigate increasing dropout rate. For the 500 pg and 60 pg conditions, one-contributor hypotheses were compared, B under Hp and X under Hd, while for the 15 pg condition dropin was also modelled under both hypotheses ( Table 3). For the two-contributor investigation eight replicates of each of two conditions were created (Table 2). The major and minor contributors were reversed between conditions, with an increased DNA contribution from the minor. These samples were amplified and analysed as described previously. Two-contributor hypotheses were compared, with each of A and C in turn playing the role of Q, while the other contributor was treated as unknown. Additionally one-contributor-plus-dropin hypotheses AUY-922 in vitro were compared, with only the major contributor playing the role of Q (Table 3). For the three-contributor investigation eight replicates of each of four conditions were created (Table 2). The conditions were created to investigate
different profiling protocols. The Phase 1 and Phase 2 conditions are post-PCR purification protocols designed to enhance the sensitivity of detection of the standard protocol [12], and both involve concentrating the post-PCR product using an Amicon® PCR microcon unit according to the manufacturer’s recommendations. Phase 1 enhancement increases the amount of formamide in the mixture compared to the manufacturer’s recommendations, while Phase 2 enhancement increases the amount of DNA, formamide and ROX compared to Phase 1. For all four conditions (30 cycles, 28 cycles, Phase 1, and Phase 2), three-contributor 3-mercaptopyruvate sulfurtransferase hypotheses were compared, with A playing the role of Q and the other contributors
treated as unknown (Table 3). Dropin was not modelled under either hypothesis, although dropin was included in the simulations. This reflects a realistic challenge for few replicates with multiple contributors, whereby any dropin alleles may be wrongly attributed to one of the contributors. However the incorrect model will lead to deterioration of inferences for larger numbers of replicates. All of the conditions that we now describe were simulated in eight replicates, with the whole simulation being performed five times. Initially a number of single-contributor CSPs were simulated using the profile of individual B. The first condition investigated was a “perfect match”, in which all eight replicates generated exactly the profile of B. Next, we introduced mild dropout (Pr(D) = 0.4) and severe dropout (Pr(D) = 0.8) of the alleles of B, in each case with dropins included at rate Pr(C) = 0.05 (at most one dropin per locus per replicate).