If the lab did not retain a sample of the DNA it tested, that's a huge problem - it's retained in vials and all labs know to keep vials for the defense.
Sometimes the defense is upset that "not enough" testing was done or not enough vials available. In that case, Judge would probably order the sheath to be sent to an independent lab, to make new vials of samples, and then send one third to the defense, one third to the prosecution and keep a reference sample (in case the defense squawks again). At some point, the defense may try to claim that the DNA on the sheath is inconclusive, but probably will dig itself a hole by trying to convince a jury of that.
I think it will turn out that BK's DNA will turn up other places (and *has* turned up other places) at the crime scene. A PCA is supposed to tap into the original crime scene analysis and the original warrants (this is a probable cause for the arrest of a particular person - there were at least 50 other warrants served very soon after the murders, I'm sure data is still coming in).
As to your last question, the answer is "sort of." There's one company that has patented a DNA-human body reconstruction software program that draws on a host of SNP markers (specific alleles of genes) to reconstruct face and other characteristics. The results are controversial because we aren't sure we even know *all* of the genes involved in, say, eye color or hair color. Some genes get turned off, some get turned on; some of this is done by other genes; some genes go "rogue"; the genes get turned on and off by the environment, but also by various processes in fetal and childhood development). That being said, here are some Parabon reconstructions where they later found the perp:
snapshot.parabon-nanolabs.com
What do you think? Very interested in opinions on this.