I have been reading and skimming through Homology, Genes and Evolutionary Innovation after seeing the book being recommended by TsviBT.

Two things I learned from this book:

The whole book is a very thorough (almost 500 pages) account of how homology in the original sense of the homology of body parts is grounded in anything. Presumably, some biologists were casting doubt on whether the homology of body parts makes any sense as a concept at all.

Homology standard example:

At the time the book was written, it was already well established that talking of homology made sense for genes. There are two different types of homology we can have for genes. Orthologs (this gene is essentially in the same location as the old gene, but has changed between species), and paralogs (this gene is a copy of the original gene, but in a different location). The whole reason why we can trace these families is because genes don't copy that often and don't change that much. Wagner is arguing the same is true of body plans. Body plans are really rigid and even more so as with the case with genes, new body parts are not developed from scratch, but by changing the logic of existing body parts or reusing the logic of existing body parts.

An example comes from looking at tusks and narwhal "horns". If new body parts readily developed, the question "What is the equivalent of the narwhal horn in humans?" would not make any sense. In practice, they are the narwhal's incisors and the same for the elephant's tusks. Different tooth types are developmentally distinguished and can develop very different forms. Meanwhile, teeth cusps don't have such individualised selection going on (which is something some biologists thought at some point, and then they checked, and it didn't properly make sense with horse evolution).

How rigid the gene networks for body plans are can be seen with the famous Hox genes:

There are numerous other examples in the book. Another core example the book brings up is the Yamanaka factors Sox2 and Oct4 that together dimerise to promote the embryonic stem cell state by promoting Nanog, which itself dimerises with the Sox2/Oct4 dimer. Because these factors are so well conserved, the Yamanaka factors work well in all mammals. I feel less surprised now that things like the Yamanaka factors exist and make it more plausible to me [supersox](https://doi.org/10.1016/j.stem.2023.11.010) might very well work in all mammals in which the Yamanaka factors work. 

The second thing I learned was about the complexity of organisms. I already knew most obvious proxies for "complexity" of an organism, like the length of the genome, don't work. What this book taught me is that there are two proxies that seem to kind of work (the number of different cell types and the number of microRNAs). Although the book mostly shows they correlate with each other, the authors already expect us to believe that the number of different cell types is a good proxy for the complexity of the organism (which I am willing to believe).