Another reason horizontal gene transfer is a big factor among microbes is that they lack dedicated germ cells. If foreign DNA gets incorporated into the genome of any cell, it will be inherited by any descendants of that cell. In contrast, in multicellular animals, any foreign DNA incorporated into the genome of a liver cell will not be inherited by anything. So, you not only have to get the foreign DNA into the nucleus, but it also needs to get into the nucleus of the right cell.

Horizontal gene transfer in complex, multicellular animals was expected to be rare. When researchers started sequencing animal genomes, they found lots of bits of viruses scattered throughout most of them. But they didn’t find many pieces of bacterial DNA. That was partly because the software that assembled the genome from individual fragments of genome sequence was made to treat bacterial sequence as contamination. That is not unreasonable, given that we were typically growing up lots of copies of the animal DNA by placing it in bacteria.

Since then, we’ve developed techniques that allow us to sequence DNA without growing lots of copies in bacteria. We’ve also got the ability to obtain sequence-extended fragments of DNA, sometimes many thousands of bases long. These “long read” DNA sequences will often cover both borders where the bacterial sequence meets the animal sequence, making clear that the bacterial version wasn’t the result of contamination.

link to open access paper https://www.pnas.org/doi/10.1073/pnas.2604240123