Date of Award

Spring 5-15-2019

Author's School

Graduate School of Arts and Sciences

Author's Department

Biology & Biomedical Sciences (Evolution, Ecology & Population Biology)

Degree Name

Doctor of Philosophy (PhD)

Degree Type



Most amino acids are encoded by multiple synonymous codons. Although alternative usage of synonymous codons does not affect the amino acid sequences of proteins, researchers have been reporting evidence for functional synonymous codon usage at the species- and gene-specific levels for over four decades. It has been shown that variations in synonymous codon usage can affect phenotypes through diverse mechanisms such as shaping translation efficiency and mRNA stability. On the other hand, the common view that cellular and organismal phenotypes are primarily determined by proteins whose functions are primarily determined by amino acid sequences, often drives the assumption that synonymous mutations are evolutionarily neutral. Consequently, this assumption has been used extensively in evolutionary biology, population genetics, and structural biology. One explanation of the apparent contradiction between the empirical findings, which indicate that synonymous mutations can affect related phenotypes, and the theoretical models, which stipulate that synonymous mutations are neutral, is that neutral synonymous mutations represent the general rule while non-neutral synonymous mutations represent the rare exceptions. In my thesis, I examined this explanation by applying computational and experimental approaches, which indicated that: 1) Non-neutral synonymous mutations significantly affect a considerable proportion of protein-coding genes; 2) Gene-specific codon usage patterns, such as the preference for a specific combination of rare codons, are possibly associated with specific gene functions, such as enhancing tissue-specific gene expression; 3) Some protein-coding genes include codon clusters whose codon usage patterns cannot be explained by selection-independent processes, and thus such codon clusters seem to serve as domains affecting protein functions. Together, these data suggest that synonymous mutations should not be a priori considered neutral. Furthermore, my studies suggest that the biochemical functions of at least some proteins are not only shaped by the constituent amino acid residues but also by codon usage biases at the gene-specific and sub-genic levels. In conclusion, my thesis work suggests that many of the commonly used approaches for analyzing the selection on protein-coding DNA sequences, which rely on the assumption that synonymous mutations are generally neutral, may generate biased results. Furthermore, my studies indicate that selection on gene-specific codon usage bias has evolved to serve diverse biological functions, which are still mostly uncharacterized.


English (en)

Chair and Committee

Yehuda Ben-Shahar

Committee Members

Yehuda Ben-Shahar, Barak Cohen, Ian Duncan, Kenneth Olsen,


Permanent URL: https://doi.org/10.7936/H0Z2-V510