If evolution leads to adaptation to our environments, then why do we still get diseases?

In many cases, disease is the price we pay for evolutionary innovations. If you like having a body made up of multiple cells, then you run the risk of getting cancer. That big brain you’ve got might lead to the development of schizophrenia. And because environments often change faster than we can evolve, evolutionary adaptations that were once useful may be obsolete or even harmful in the present day. These evolutionary trade-offs and environmental mismatches suggest that we will never ‘evolve-out’ of having diseases, but by looking for the signatures of natural selection in our genomes, we can gauge how human populations has been shaped by evolutionary pressures. This has potential to not only explain the origins of complex human diseases but also give insights into their mechanisms of action.

Sequencing the genomes of thousands of humans not only provides base-pair-level resolution of the genome, but also provides insight into population-level genetic variability. Since selection acts on our DNA, different types of selection (e.g., positive, negative, or balancing) leave different signatures in our genomes. Previous efforts have studied patterns of human genetic variation to infer different types of selection, but most have not considered these different types of selection simultaneously.

This led us to develop a computational framework to evaluate diverse modes of selection on genomic regions associated with complex genetic traits, including disease. The manuscript describing this work “Accounting for diverse evolutionary forces reveals mosaic patterns of selection on human preterm birth loci” was published on Friday July 24^th 2020 in Nature Communications. The key technical innovation in this work was to harmonize different selection metrics to the same scale and generate an empirically calculated p-value that accounts for factors that influence the expected values of these selection metrics in different parts of the genome. To illustrate the power of our novel computational evolutionary framework, we applied this method to preterm birth.

Preterm birth (PTB) is a major health concern affecting up to 15 million pregnancies worldwide and 10% of pregnancies in the United States. PTB also has serious health consequences such as increased infant mortality rates and significant morbidity. You may have friends or family who have experienced the effects of PTB. They join parents like Beyoncé & Jay-Z and Anna Faris & Chris Pratt who have publicly spoken about their preterm births. While there are some factors, like hypertension, that put women at higher risk for delivering preterm, many preterm births are spontaneous (sPTB)—occurring without warning.

We currently have a limited understanding of the triggers of labor that can lead to preterm delivery.  Family-based studies have identified a strong genetic basis for preterm birth suggesting there are specific genomic regions that contribute to disease risk. Thanks to our colleagues at the Ohio Collaborative March of Dimes Prematurity Research Center, the largest genome-wide association study (GWAS) on preterm was recently published. Using the genetic variants identified by the GWAS to influence PTB risk, we asked how selection has shaped these regions.

Across the preterm birth associated genomic regions, we uncovered the signature of a mosaic of evolutionary forces. In addition to positive and negative selection, we detected balancing selection, deep conservation, and recent population differentiation. Many of these regions with signs of selection had compelling biological roles that affect preterm birth or other adverse pregnancy outcomes. This illustrates how the evolutionary context we provide helps understand the tradeoffs underlying several of these loci.

Healthy births increase the chances that our genetic material is passed onto future generations—a necessary step for natural selection. Preterm birth puts the life of the baby at risk, and thus has the potential to be under strong natural selection. Here, we show that genetic variants that influence preterm birth risk have indeed experienced a mosaic of strong selective pressures. Reproductive biology is incredibly complex, but studying how evolution has shaped birth timing provides a powerful lens to better understand the mechanisms of reproductive health and disease. Furthermore, our computational evolutionary framework is general and can be applied to any complex trait or disease. We hope that it will enable us and other researchers to catalog the selective pressures that have shaped diverse human traits.