After death, the soft tissue of the original tooth pulp desiccated and degraded, but the mineralized shell remained intact, sealing whatever organic material had survived. When the researcher working on the specimen in the Leipzig laboratory placed a small fragment of that tooth into a mill and ground it to powder — an irreversible act that consumed part of the specimen in the process of extracting information from it — they were accessing organic material that had not been exposed to the open environment since the death of the individual.
The DNA extracted from that powder was fragmented beyond recognition as a coherent genome. It existed as billions of short, chemically damaged pieces, interspersed with the DNA of thousands of other organisms that had colonized the cave sediment, the tooth surface, and the laboratory environment over the decades and centuries since excavation. The extraction protocols isolated the hominin DNA from this background by targeting sequences known to be present in hominin genomes and enriching them through hybridization capture.
What the sequencer produced was not a readable genome but a set of short sequence reads — hundreds of millions of them, each covering perhaps 50 to 100 base pairs of the original genome. The computational pipeline assembled these reads against a reference Neanderthal genome, identified where each read mapped, counted how many reads covered each genomic position, and called genotypes where coverage was sufficient. For many positions, coverage was insufficient for confident calls; for others, it was adequate; for a subset of positions critical to the analyses, targeted capture had produced sufficient coverage to make the kinship, sex, and mitochondrial analyses work.
