Evolution via the gradual accumulation of mutations

According to James Shapiro (2011, Evolution: A View from the 21st Century, p. 95):

The 2001 Nature report of the draft human genome contained two important figures illustrating what genome sequencing had taught us about protein evolution [382]. Using transcription factors and chromatin binding proteins as examples, the figures showed that these classes of proteins did not evolve one amino acid at a time [emphasis mine]. Instead, the two classes of protein "shuffled" and "accreted" copies of functional protein segments called domains as eukaryotes progressed from yeast through nematode worms and Drosophila fruit flies to mice and human beings. In other words, proteins diversify through a process of acquiring, amplifying, and rearranging coding sequences for subprotein structures that may be dozens or hundreds of amino acids in length.

References

382. Lander, E.S. et al. Initial sequencing and analysis of the human genome. Nature 409, 860-921 (2001).