Evolution Update

Evolution Update

How to Predict the Future by Looking at the Past

Samy Lafin February 22, 2015

Scientists analyze the evolutionary past of organisms to try to predict their evolutionary future.

Evolution – the changes observed in organisms over time – is a fact. If we agree that each generation carries with them genotypical and phenotypical differences from the previous generations (known in the scientific community as descent with modification), then it is hard to deny that over millions and even billions of years that organisms could have evolved into the various life forms we see today.

Researchers are able to study different lines of descent, and they organize them into phylogenetic trees. Phylogenetic trees are similar to huge family trees, indicating where common ancestors are, where branching off could have occurred, and indicating relatedness of different individuals or species.

In one study, the researchers organized phylogenetic trees, and instead of using them to look at the past, used them to help predict the future. Why would this be useful? Well, did you get your flu shot this year? If you did, you may know that this year’s (2014) batch of flu vaccines wasn’t particularly effective in helping immunize people to the flu. Each year, you have to get a new flu shot because the strains of flu change.

Influenza is a virus, which – at the most basic level – is genetic material (DNA) covered by a protein coat. How a virus infects you is to inject its DNA into your cell’s DNA, and the viral DNA takes over the cell. Instead of your cells doing what they are supposed to do, they turn into virus factories. When the cell has made enough new viruses, the cell bursts, releasing new viruses into the body to go and infect new cells. In order to make new viruses, the viral DNA must be replicated over and over again. We all know that when we make copy after copy of something, there tends to be a mistake made here or there. When a molecule of DNA copies with mistakes, we call those mistakes mutations. As these mutations accumulate, they can either cause no change in how the virus is made or they can cause the virus to change slightly. The flu is a highly prolific disease, and each year – due to all the mutations that have occurred – there are new strains that we must immunize ourselves to, resulting in your yearly flu shot.

However, the shots aren’t always effective, as we have seen this year. The reason is that scientists have to guess as to which strains of the flu will be most prevalent. Knowing this, the goal of the researchers was to use these phylogenetic trees to make predictions. The trees were used to examine the fitness – the ability of the virus to maintain/increase its numbers due to reproduction – of the influenza strains. Those strains with the highest fitness would then be those to vaccinate against, as those would be the strains predicted to be the most prolific.

The researchers hypothesized that evolution happens due to the accumulation of these mutations without a lot of external/environmental pressure. In order to test their hypothesis, the researchers applied a mathematical algorithm (a really detailed equation) to phylogenetic trees in order to predict which individuals would have the highest fitness. They made a few assumptions before they began their research. One assumption was that the fitness levels of each sample were heritable, meaning the fitness changes could be passed through the DNA. They also assumed that there are certain patterns that result when using phylogenetic trees, and that these patterns could be used as a model to predict what will happen.

That last part might sound a little dicey to you, but this sort of science happens all the time. How do weather forecasters make predictions about what the weather will be like in the future? How do economists infer what the market will do in the coming months? They use models – just like our researchers are doing. Meteorologists look at several different models based on current weather and trends to understand how the weather systems will move and change. Economists use mathematical models to predict what the market will be doing, based on current trends. Our researchers are just applying an old technique to a new area.

The researchers looked at the shapes of the resulting phylogenetic trees, certain “neighborhoods” of the trees, distance between “children”, and applied their algorithm to understand what will happen to those individuals with the highest fitness. They worked under the assumption that the offspring with the highest fitness will be the ones to survive and produce the next round of offspring. Also, the researchers tuned their algorithm to allow for a simple model of evolution – change over time – and allowed for certain beneficial mutations in order to predict the next rounds of offspring.

Using their phylogenetic trees, the researchers determined fitness by the amount of branching in a lineage. If the lineage had a lot of branching, the researchers deemed these the lineages with high fitness. They were producing a lot of offspring. Remember, we stated earlier that the more copies you make, the more mistakes you make. Those mistakes would accumulate and result in the branching. Lineages with long periods of no branching were deemed to have low fitness. With this branching data, the researchers were able to determine a local branching index(LBI), which looked at “total tree length” and the distance from the common ancestor. They compared the LBI to the more complex selection-based diffusion (SBD) model, which was a mathematical approach to predicting offspring. The SBD model “assumes evolution proceeds via accumulation of many small effect mutations.” In other words, if we can estimate the number of mutations that occur each generation, we can anticipate what the genetic makeup of the offspring will be and use those results to predict what we will see in future generations. The data showed both models yielded similar results.

The researchers took their models, once they had tested them enough to understand their inner workings, and applied them to influenza data from 1995-2013. They were able to accurately predict influenza strains for 9 of those years, and only totally failed in 3 of the 19 years. For the remainder of the years, they deemed the results of being “intermediate accuracy”. Overall, “the sequence with the highest inferred fitness tends to be a close match to the progenitor of future populations.” In layman’s terms, those individuals that were able to produce the most offspring (highest inferred fitness) tended to be the main contributor to future generations. What that means for us, is that if we can accurately predict which strain of influenza will be the most virulent and prolific, then we can predict which strains to immunize against.

At the end of the study, the researchers brought up some ideas to improve upon their method. First, they noted that influenza populations can vary in terms of fitness (which is why some years the flu is “bad”, and other years the flu doesn’t seem as virulent). Then, they stated that predictions could get better if we can apply information regarding the shape of the phylogenetic trees, we might be able to enhance the models to obtain better predictions. It was also noted that we could combine information regarding tree shape “with antigenic information,…biophysical and structural knowledge,…patterns of past evolution,…and plausible geographic sources” to further define the models and increase our chances of success at predicting offspring.

So, for those looking for a quick summary, looking at the evolutionary history of genetic traits can determine patterns that can be used to predict the evolutionary future.

Recent Articles

"Why Do Those Flowers Look like Bugs? Or, on the Evolution of Orchids."
A large group of flowering plants, commonly known as Orchids, often have flowers whose shape coincides with that of their insect pollinators. Recent research has shown how this uncanny flower morphology is guided by evolutionary selection.

"How Plants Maintain a Low-Sodium Diet Without Advice from Their Doctors"
Salt tolerance is a critical stress response in many plants and is controlled by a wide variety of interacting genes. Researchers studying sodium transporters in trees from high-salinity environments have characterized the evolution of these genes and determined that they are under strong positive selection in salty soils.

"Evolutionary History of a Widespread, Recently Diverged Antioxidant Enzyme in a Pig Pathogen"
Peroxiredoxins are proteins conserved across all domains of life that protect cells against the threat of reactive oxygen species. Researchers have recently characterized the evolutionary history of an essential peroxiredoxin gene from a common livestock pathogen.

"A New Class of Antibiotics Less Susceptible to Evolutionary-Driven Resistance Development"
Pathogenic bacteria are evolving resistance to our antibiotics at an alarming rate, however, scientists have recently discovered a molecule that may help combat these microscopic killers.