Collection of evidence is also another likely place to make a mistake. The issue is that a non-pristine sample taken from a public place (like a crime scene) may have multiple DNA profiles represented in it. I'm not highly familiar with the technology but I've read that techniques like PCR that "amplify" the DNA collected from the original sample by synthesizing copies of the DNA can further exacerbate the issue.

> One of the biggest strengths of PCR(e) for DNA typing is the degree to which DNA can be amplified. Starting with a single DNA molecule, millions or billions of DNA molecules can be synthesized after 32 cycles of amplification. This level of sensitivity allows scientists to extract and amplify DNA from minute or degraded samples and obtain useful DNA profiles. In this context, the sensitive nature of PCR works in a lab's favor, but it can cause problems if great care is not taken to avoid contaminating the reaction with exogenous DNA.

https://www.promega.com/~/media/Files/Resources/Profiles%20I...

> Because extremely small samples of DNA can be used as evidence, greater attention to contamination issues is necessary when identifying, collecting, and preserving DNA evidence. DNA evidence can be contaminated when DNA from another source gets mixed with DNA relevant to the case.

https://www.ncjrs.gov/nij/DNAbro/evi.html