Fig. 5 shows the information block for a candidate allele of locus Penta E. It is the only erroneous sequence that was not automatically filtered by the 10% default threshold. The information supports that this candidate allele should be disregarded. The putative allele length is one STR repeat unit smaller than the high abundant

(47.40%) sequence with index 6, indicating that it might be stutter. Apart from this stutter there are no other sequence differences (Ist relation degree). Furthermore, the clean flank percentage is rather low (59.5%), indicating possible low quality Pexidartinib sequences. An unexpected strand distribution of 100% implies that there are no complementary reads supporting the presence of this allele candidate. Removing this allele candidate Panobinostat is accomplished by unchecking the “in profile” check-box. After selecting the “Length-based analysis” check-box, all allele candidates are displayed proportionally, according to their actual length within the locus, as shown in Fig. 3. For each locus, the x-axis is adjusted to show the locus length starting from the shortest allele and ending at the longest allele. The threshold bar is no longer displayed because allele

candidates with the same length are now stacked on top of each other, which creates one bar that shows the total abundance of all alleles with the same length within each locus. This representation resembles a CE profile. The example of the allele candidate in Fig. 5 now visually looks like a CE stutter peak based on the relative length and abundance difference as compared to the true Tau-protein kinase allele. After reviewing the profile by setting the threshold to an appropriate value, and removing allele candidates of poor quality, pressing the “Make profile” button yields the final profile. This profile can then be used to query databases or compare to the profile of a sample of interest. Fig. 6 shows the final profile for sample 9947A_S1. Using the threshold of 10%, it has

one Penta E allele 13 that is undetected relative to the known genotype (Table A.1). This allele is present in the data at an abundance of 8.85% and its corresponding green bar can be seen clearly in Fig. 3. The sub-optimal results of the pentanucleotide loci, Penta D and Penta E, were previously discussed in detail [9]. We show how an MPS data-set can be analyzed using an easy-to-use graphical user interface, requiring a limited number of parameters and almost no bioinformatics expertise. The interactive visual representation of the results shows additional information when hovering over the alleles, allowing for in-depth analysis of the underlying sequences and the related statistics. For clarity of explanation we chose to display and discuss the analysis of a single contributor sample, but the MyFLq framework equally works on mixtures because no assumptions on mixture composition are made to perform the analysis.