In another seeming step towards Jurassic Park, two groups of researchers recovered full-length mitochondrial DNA sequences from 22,000 to 44,000 year-old bones of extinct European and North American bears. Full-length mtDNA has been recovered from similarly ancient specimens, but in those cases frozen tissues preserved in permafrost were used. Both groups used specialized PCR protocols employing several hundred primer pairs designed to recover short fragments, rather than one of the newer sequencing technologies, demonstrating the continued power of DNA amplification.
In 28 july 2008 BMC Evol Biol Proc a group of 18 researchers led by Johannes Krause, Max Planck Institute, Germany, recovered full-length mtDNA from a 44,000 year old Ursus spelaeus (European cave bear) bone found in an Austrian cave, and from a 22,000 year-old skull of Arcdotus simus (American giant short-faced bear) from Eldorado Creek, Canada. In 11 november 2008 Proc Natl Acad Sci USA, 14 researchers led by Jean-Marc Elalouf, Institute de Biologie et Technologies de Saclay, France, report full-length U. spelaeus mitochondrial genome from a 32,000 year-old bone from the legendary Chauvet-Pont d’Arc Cave, home to the oldest rock art pictures ever found.
If we found a bone from one of these extinct bears in our backyard, could it be identified by its COI barcode? Submitting the long-ago bears’ COI barcode region sequences (positions 48 to 705) to BOLD ID engine flags both species as not in database, with a NJ tree similar to that created by full-length genomes (ie the extinct U. spelaeus is sister to U. arctos (Brown bear) and U. maritimus (Polar bear), and extinct Arcdotus simus is 
sister to Tremarctos ornatus (Spectacled bear). Of course it would be difficult to recover a full-length sequence–what about the 130 base pair “mini barcode” proposed for broad-scale biodiversity analysis? This is within the size range(ie < 180 bp) that Elalouf and colleagues report best for recovery of ancient DNA. Remarkably, A. simus mini-barcode submitted to BOLD ID engine gives NJ tree correctly showing T. ornatus as its sister species and U. spelaeus mini-barcode correctly picks out U. arctos and U. maritimus as most closely-related species.
Recovering DNA from ancient bones leads to CSI-like thoughts of where else we might usefully recover DNA for species identification. DNA has been recovered from naturally shed feathers, flakes of seal skin at breathing holes in polar ice, hair and saliva left by predators of sheep, bird faeces, and, turning to world of commerce, ancient and modern processed leather goods (Long 2007). I look forward to analyses of the many processed foods with what is currently an unverifiable “list of ingredients.”