Digging Data

Mutations – changes in the genetic sequence of DNA or RNA – are the raw material for evolution.  Natural selection, genetic drift, and other evolutionary processes act on genetic variation – and that genetic variation starts with mutation.  Even if a genetic variant is introduced to a population through migration, ultimately, that variant got its start as a mutation. So understanding where, when, why, and how often mutations occur is key in understanding how evolution happens. Today, quick and inexpensive DNA sequencing means biologists can take a closer look than ever before at the process of mutation. What they are learning is sometimes surprising.

In humans, each baby has around 70 brand new or “de novo” mutations. De novo mutations occur in the reproductive cells of parents and are passed on to the child.  Evidence suggests that most de novo mutations in a child come from the sperm that helped create that child, and relatively few mutations come from the egg. Biologists thought this made sense. In humans, beginning at puberty, the cells that produce sperm divide (and copy their DNA) throughout adulthood, leading to vast numbers of sperm. In contrast, in a person with ovaries, all the DNA copying leading up to egg production is completed before that person is even born.^[1] The cells that produce sperm just go through many more cycles of DNA replication and cell division than do the cells that produce eggs.^[2] If most mutations happen because a cell makes an error when it copies its DNA, producing a new DNA strand that differs slightly in sequence compared to the original, then we’d expect sperm to be the source of most new mutations. All those cell divisions in the cells that eventually lead to sperm provide many opportunities for copying mistakes to occur.

Biologists Felix Wu, Alva Strand, Laura Cox, Carole Ober, Jeffrey Wall, Priya Moorjani, and Molly Przeworski work at different universities and research centers, but they all wanted to know, is it really the case that most new mutations in humans are caused by copying errors when cells divide? Some evidence had already suggested that copying errors are not the whole story when it comes to heritable mutations.

Hypothesis

The research team set out to test the hypothesis that most de novo mutations occur because of copying errors. They looked at what percent of new mutations in a baby come from DNA carried by sperm, as opposed to from DNA carried by the egg; we’ll call this “sperm bias” in new mutations. If new mutations mostly come from copying errors, then the number of mutations contributed by eggs should not be much affected by the age of the person contributing the egg (since, in primates, all the DNA copying that leads up to egg production occurs before that future parent is even born).  However, the number of mutations contributed by the sperm should increase with the age of the genetic father (since DNA copying and cell divisions leading to sperm are ongoing from puberty throughout adulthood). Overall, this means that if the hypothesis were true, we’d expect to see that species with younger genetic fathers (and a shorter interval between puberty and fatherhood) should have less sperm bias in new mutations than species with older genetic fathers (and a longer interval between puberty and fatherhood).

Data

The team decided to compare mutations in humans to mutations in olive baboons (Papio anubis). The two primates are closely related, but male baboons go through puberty at age 6 and reproduce at age 10 on average, while humans that produce sperm typically start puberty around age 13 and don’t reproduce until age 32 on average. This means that the cells that produce baboon sperm go through about 4 years’ worth of DNA replication and cell division before a sperm leads to offspring, while the cells that produce human sperm go through about 19 years’ worth of DNA replication and cell division before a sperm leads to a child!  That’s a big difference, providing lots more opportunity for mutations caused by DNA copying errors to accumulate in the sperm-producing cells of humans compared to those of baboons. To get data on mutations, the team sequenced the genomes of three generations of three different human families (26 people total) and three generations of two baboon families (29 baboons total).

They focused on tallying up selected spots in the genome where the DNA sequences of the parents were the same, but their child’s genetic sequence was different from that of both parents.^[3] The only way to explain this outcome is if one of the parents contributed a brand new mutation to the child.  Then the researchers used additional sequence information to determine which parent (mother or father) had contributed each mutation.^[4] This is what they found:

In these graphs, each offspring in which new mutations were identified is represented by a circle and a triangle.  The circle indicates the number of mutations contributed by the sperm and the triangle indicates the number contributed by the egg. The lines show the overall relationship between age at conception (on the x-axis) and mutations (on the y-axis). The shading around these regression lines indicates the 95% confidence intervals.

Right away, we can observe a few things.  First, the blue circles are always positioned above the red triangles, indicating that sperm contribute more mutations than do eggs. That’s just as we’d expect based on other research. Second, in humans, older genetic parents do seem to pass on more mutations to offspring, as shown by the upward slope of the regression lines. In baboons, the mothers’ age is associated with more mutations (the line slopes upwards), while the regression line for fathers is nearly flat – suggesting that older fathers don’t contribute more mutations. This result goes against previous findings about mutations and age. But if you look at the male baboon data closely, you’ll notice that the blue confidence interval is wider than the others. This means that the data vary a lot, making it hard to detect patterns without a really large sample. So it is possible that baboon fathers, like humans, do pass on more mutations to their offspring as they age, but there were not enough data in this experiment to detect the pattern.

The researchers used their data to calculate the proportion of all new mutations that came from the father (i.e., were paternal), as shown in the graph below.  The vertical lines represent the 95% confidence intervals for each point.  A chi-squared test showed that the two values are not significantly different (p = 0.91). Even though the baboons were much younger than the humans when they reproduced (and the cells that produce their sperm have been through many fewer rounds of DNA replication and cell division), the degree of sperm bias in the two species is similar! That does not match the expectation generated by the hypothesis that most new mutations come from copying errors.

If not copying errors, what is the source of the new mutations that primate offspring are born with? The research team suspects that sloppy DNA fixes play a role. DNA, including the DNA within sperm and eggs, is easily damaged.  And when it is, a cell does its best to put the strand back together perfectly.  But sometimes it mistakenly substitutes one genetic letter for another, generating a mutation.  The idea that errors in DNA repair are an important source of mutations that matter for evolution fits with many lines of evidence – including the evidence described here that the number of new mutations increases with the age of both genetic parents. After all, we might expect the chances of DNA damage and sloppy repair to be fairly constant over time, so as the cells that lead to eggs and sperm get older, we’d expect them to accumulate more mutations, which are passed on to daughter cells and ultimately, to the eggs and sperm themselves. But exactly how important this process is, as well as how it might vary between the sexes and across species, has yet to be worked out. Furthermore, the idea that DNA damage and poor repair are the main source of de novo mutations does not help us understand why we observe sperm bias in the first place. As our genetic techniques advance further, biologists like Felix, Molly, and their colleagues will have even more data to home in on the processes that lead to mutations, tracing genetic variation back to its original source.

Stepping into science

The research team was led by Felix Wu, a graduate student at Columbia University in New York City, and Molly Przeworski, a professor there. Molly did grow up in a large city and spent most of her leisure time reading novels. She was utterly uninterested in math or science—her least favorite topic in school was biology. But a great teacher drew her into math in college and after a series of false starts, she discovered genetics and evolutionary biology.