Re: Riman et al. Examining performance and likelihood ratios for two likelihood ratio systems using the PROVEDIt dataset

The editors,
Recently Riman et al. [1] published a comparison of the two PG software - STRmix™ V2.6 and EuroForMix V2.1.0 (EFM) - using STR profiles generated using the GlobalFiler kit from the open source PROVEDIt dataset [2]. Whilst both software perform well, if used properly, there are a number of technical deficiencies to consider when assessing different results from analyses of the same DNA profile with each software. Some of the technical faults are sufficiently significant to render the data valueless. Others may have minimal consequences. We discuss these faults and their impact below. We have previously written to the authors offering assistance with technical aspects of the use of STRmix™ specifically to ensure proper use and understanding of a tool of which we have substantial experience. This offer was not accepted; perhaps the lack of communication was to maintain some independence by the authors with their study. While we laud independence and avoidance of conflict of interest, there is a balance and seeking some input could have overcome notable and obvious design flaws of their current study.