I'm not sure that I buy that 90% number. The percentage of our DNA that is composed of endogenous retroviral material
is around 5-8%
. ERVs are horizontal gene transfers that occur in germ-line
cells, such as sperm, ova, and all of the cells in their ancestry back to the original zygote for that individual. Genetic changes to these cells (and only
these cells) will be passed down to future generations.
Now, it's true that ERVs are not the only type of viral DNA that an individual may have in their cells. Any infection of a somatic (non-germ-line) cell by the appropriate type of virus since the individual's conception will lead to chimeric DNA in some part of the body. For example, well over 90% of American adults have had some form of herpes
infection during their lives, such as chicken pox or herpes simplex. This becomes a permanent addition to the DNA in the infected portions of the body, but it is NOT passed down to offspring.
The reason that your 90% figure doesn't pass the sniff test is because it would mean that more than 80% of the DNA in an individual's body would be acquired AFTER birth. If this were true, then wouldn't we expect to see huge, obvious differences between individuals throughout the entire genome? This is definitely not what we see when we sequence DNA. After all, which diseases an individual contracts, when they contract them, and in what order is essentially never the same. Hell, the difference between a human and a chimpanzee's genome is only about 4%
. The difference between individual humans is far smaller than that, so it seems likely that only a small (probably 1%) percentage of an individuals genome is made up of viral material obtained since birth. This passes the sniff test as well; you'd expect the genetic insertions that have accumulated over millions upon millions of years in germ-line cells to far outweigh the horizontal gene transfers that happen within a single individual's lifetime.