Working for a company doing rapid PCR tests I was briefly put on a project with the FBI to speed up DNA forensics, and saw they had multiple problems with their testing. The key problems being that there were too many false negatives and too many false positives. I explained this, presented solutions, and was promptly removed from the team. Here's why DNA forensics sucks:
1) too little data: they verify ID based on matching 6-8 short segments of code, out of a massive codebase. Imagine identifying a dictionary using partial definitions of six words.
2) too many subjects: this method works 50% of the time when the database has less than 100,000 samples in it, but other times IDs 0 to 10 individuals. Put a billion samples in the database and there will be many more hits.
3) genetic drift: Our DNA evolves over time, and particular organs will re-organize so that e.g. a hair sample won't match a blood sample in the same individual. Or a hair sample 20 years later won't match the earlier one.
The above cited law-enforcement officials are basically ID-ing their target as "black man in blue car, somewhere near Detroit", and running with it..
That said, DNA/RNA is the coolest codebase I've worked with: bombproof syntax, error-correction, data-retention, execution efficiency, etc. and that's just the 1% that made sense!