Techniques used to infer pathways of protein evolution found unreliable

A key assumption that biologists have relied on widely over the past quarter-century in studying the evolution of protein molecules is “highly questionable,” according to an article published in the November issue of BioScience.

The article, by Shozo Yokoyama, a vision researcher at Emory University, summarizes experimental work that involved creating and measuring the properties of dozens of reconstructed ancestral versions of visual pigments found in the eyes of vertebrates, including humans, as well as deliberately altered variants. Yokoyama concludes that the studies he assessed “cast serious doubt” on the “fundamental principle of molecular adaptation,” the bedrock of thousands of published papers based on reconstructions of evolutionary changes in a wide range of proteins. The statistical tools used by such studies are “not reliable,” he writes.

In attempting to understand how proteins and their properties might have changed over time, biologists have typically made simplifying assumptions. One is that a known change at a particular spot in a protein would affect the properties of ancestral and modern forms of proteins in similar ways. That simplification makes it relatively easy for computers to infer the likely evolutionary paths that led to the forms of the proteins found in modern organisms—for example, the visual pigments found in deep-sea fishes (which live with no ultraviolet light) and the different pigments found in shallow-water fishes.

Yokoyama tested the assumptions by making the hypothesized ancestral pigments and variants of them that might have been produced by mutation, then accurately measuring their properties. The results were disturbing: the properties of related versions of proteins would often change in very different ways when the same mutation was introduced into each. Consequently, standard computational and statistical methods would rarely have identified the experimentally supported evolutionary pathway. Yokoyama expresses the hope that other researchers will start to make and test the properties of reconstructed ancestral proteins to evaluate hypotheses about their evolution, rather than relying on computational approaches.

BioScience, published monthly, is the journal of the American Institute of Biological Sciences (AIBS; www.aibs.org). BioScience is a forum for integrating the life sciences that publishes commentary and peer-reviewed articles. The journal has been published since 1964. AIBS is a meta-level organization for professional scientific societies and organizations that are involved with biology. It represents nearly 160 member societies and organizations. The article by Yokoyama can be accessed ahead of print at www.aibs.org/bioscience-press-releases/ until early November.

The complete list of peer-reviewed articles in the November, 2012 issue of BioScience is as follows. These are published online ahead of print from 11 October.

Synthesis of Experimental Molecular Biology and Evolutionary Biology: An Example from the World of Vision.

Shozo Yokoyama

Sentinels of Ecological Processes: The Case of the Northern Flying Squirrel.
Winston P. Smith
What Is Conservation Science?
Peter Kareiva and Michelle Marvier
A State-Based National Network for Effective Wildlife Conservation.
Vicky J. Meretsky, Lynn A. Maguire, Frank W. Davis, David M. Stoms, J. Michael Scott, Dennis Figg, Dale D. Goble, Brad Griffith, Scott E. Henke, Jacqueline Vaughn, and Steven L. Yaffee
A Global System for Monitoring Ecosystem Service Change.
Heather Tallis, Hal Mooney, Sandy Andelman, Patricia Balvanera, Wolfgang Cramer, Danny Karp, Stephen Polasky, Belinda Reyers, Taylor Ricketts, Steve Running, Kirsten Thonicke, Britta Tietjen, and Ariane Walz