I'd almost forgotten about the forensics ILOR. It sounds like a productive meeting, even if the legal go-ahead hadn't been given at that point.
Just a few thoughts for the pot...
From memory, only one hair in 5A was ever positively identified and that was just Gerry's. Three others had partial roots, but not enough to extract a nuclear DNA profile from. I'd read about attempts to extract nuclear DNA from the top of rootless hairs in the past couple of years, but that seemed fairly hit and miss.
At the time, mtDNA forensic analysis seems to have been limited to establishing a haplotype, which seems about as useful as establishing a dialling code - i.e., a bit better than nothing. It may have been due to legal wrangles over ethical and privacy issues, as medical research pinpointing faulty or mutant genes in the mitochondria was well advanced way before Madeleine disappeared, so the technical capability existed.
I'm also curious about the curtains (and the bedding). In an ideal world, the curtains in the kids' bedroom and behind the sliding door should have been bagged as the capablities of extracting forensic information from cloth has apparently also made progress. I suppose there could also be a small miracle and they actually were, but it's not logged in the files.
If all of these, plus the identities of phone accounts / cybercafe users / people who had access to or had made duplicate keys could be established, there might be a few red flags. Particularly if someone known to have been in the vicinity denied ever having been in there. And, conversely, eliminate a number of people, thus narrowing down potential people of interest.
