The pattern of inert retroviruses in our DNA too closely matches that of chimpanzees to be explained by uncommon descent (endogenous retroviruses, as they are called). We don't share just a few ERVs with chimpanzees, we share most of them (and there are almost a hundred thousand of them). We know that these sequences were put there by retroviruses and not designed in because (1) they have all the features of a viral DNA insertion event (including target-site duplications, long terminal repeats, gag, pol and pro regions), (2) viral products are sometimes produced by ERVS and (3) scientists have been able to ressurect at least one such ERV back into an actual virus (
the Phoenix virus).
The location in the genome where retroviruses insert themselves is highly random. This is true even when they insert themselves into different cells of the same species (as HIV studies have shown), let alone different species. Although retroviruses do have preferred regions where they insert themselves, they do not have prefered loci (i.e.
exact positions within DNA). If these ERVs had inserted themselves separately in chimp and human lineages, then they would be in mostly different spots (we are talking about billions of potential locations for the random insertion of ~98,000 ERVs). ERVs are a class of genetic elements called indels (short for "insertion/deletion"). Not all indels are ERVs and ERVs make up about 10% of total indel content. The Chimpanzee Genome Project has shown that human and chimp indel sequences are 97% similar. Even if we assumed that the entirety of the 3% difference took the form of ERVs in different loci (which would be 30% of the ERVs), that would still mean that at least 70% of the 98,000 ERVs are in the same spot. The study also found less than 100 ERVs that were specific to humans and less than 300 specific to chimps, suggesting that the actual similarity between our ERVs is in excess of 99%. You simply do not get numbers like that from random insertions in separate lineages.
My source.