ESR commentary on California vs. Dominguez

A controversial murder case which involved evidence produced through the use of STRmix software recently ended in a plea bargain agreement in which with the defendant pleaded guilty to a single charge of manslaughter in the 2008 death of a 15-year-old.

 
Reporting on the case, a reporter for the San Diego Union Tribune wrote, “When [defense attorney] Speredelozzi sought to have an expert examine the source code to determine how the program is working, he was rebuffed. The company [Institute of Environmental Science and Research, or ESR] said he would have to sign a restrictive nondisclosure agreement first, claiming the code was a privileged trade secret it could protect.”

In point of fact, Speredelozzi was not rebuffed and the nondisclosure he was asked to sign is not restrictive. ESR (which is a government institute, not a private company as characterized in the Tribune article) has the utmost respect for Sixth Amendment rights. As a result, ESR voluntarily provides access not only to the source code for its STRmix software, but also to a time-limited version of the software, developmental validation records, user’s manuals, and extended output.

While ESR does ask prospective users to sign a nondisclosure agreement, that NDA is, by all counts, more liberal than the terms offered by virtually any other commercial software and is simply designed to safeguard ESR’s intellectual property rights 

The article was also inaccurate on multiple other counts. Specifically, it failed to note that:

• STRmix source code has already been reviewed four times by the same defense expert.

• All of the algorithms for STRmix previously have been published in peer reviewed literature – to the extent that partial clones of the software now exist.

• STRmix has been extensively validated to ensure robust scrutiny and acceptance by the scientific community.

• ESR negotiated in good faith with the defense in the case cited.

Bottom line with respect to this case: Failure to reach an agreement is not the fault of ESR. ESR’s efforts to reach agreement were both extensive and costly. Given the resulting plea bargain, ESR can only assume that the defense was not genuine in its interest in reviewing the source code, and potentially was more interested in giving the impression that ESR was being uncooperative.

Probabilistic genotyping and STRmix continue to be generally accepted by the forensic community and represent a significant improvement over older methods of DNA profile interpretation. Standards for probabilistic genotyping software have been published by professional bodies covering the validation and use of such software. These include SWGDAM and the UK Forensic Science Regulator (FSR).
 

More importantly, probabilistic genotyping methods are much more likely to correctly exclude individuals who are not present in a DNA mixture and, therefore, to be beneficial in proving innocence. STRmix and other probabilistic genotyping software have been used repeatedly to overturn wrongful convictions that were based on less powerful interpretation methods.