What is a crossover anyway, and why do I, as a genetic genealogist, care?
A crossover on a chromosome is where the chromosome is cut and the DNA from two different ancestors is spliced together during meiosis as the DNA of the offspring is created when half of the DNA of the two parents combines.
Identifying crossover locations, and who the DNA that we received came from is the first step in identifying the ancestor further back in our tree that contributed that segment of DNA to us.
Crossovers are easier to see than conceptualize.
The crossover is the location on each chromosome where the orange and black DNA butt up against each other – like a splice or seam.
In this example, utilizing the Family Tree DNA chromosome browser, the DNA of a grandchild is compared to the DNA of a grandparent. The grandchild received exactly 50 percent of her father’s DNA, but only the average of 25% of the DNA of each of her 4 grandparents. Comparing this child’s DNA to one grandmother shows that she inherited about half of this grandmother’s DNA – the other half belonging to the spousal grandfather.
- The orange segments above show the locations where the grandchild matches the grandmother.
- The black sections (with the exception of the very tips of the chromosomes) show locations where the grandchild does not match the grandmother, so by definition, the grandchild must match the grandfather in those black locations (except chromosome tips).
- The crossover location is the dividing line between the orange and black. Please note that the ends of chromosomes are notoriously difficult and inconsistent, so I tend to ignore what appear to be crossovers at the tips of chromosomes unless I can prove one way or the other. Of the 22 chromosomes, 16 have at least one black tip. In some cases, like chromosome 16, you can’t tell since the entire chromosome is black.
- Ignore the grey areas – those regions are untested because they are SNP poor.
We know that the grandchild has her grandmother’s entire X chromosome, because the parent is a male who only inherited an X chromosome from his mother, so that’s all he had to give his daughter. The tips of the X chromosome are black, showing that the area is not matching the mother, so that region is unstable and not reported.
It’s also interesting to note that in 6 cases, other than the X chromosome, the entire chromosome is passed intact from grandparent to grandchild; chromosomes 4, 11, 16, 20, 21 and 22.
Twenty-six crossovers occurred between mother and son, at 5cM. This was determined by comparing the DNA of mother to son in order to ascertain the actual beginning and end of the chromosome matching region, which tells me whether the black tips are or are not crossovers by comparing the grandchild’s DNA to the grandmother.
For more about this, you might want to read Concepts – Segment Survival – Three and Four Generation Phasing.
Before going on, let’s look at what a match between a parent and child looks like, and why.
If you’re wondering why I showed a match between a grandchild and a grandparent, above, instead of showing a match between a child and a parent, the chromosome browser below provides the answer.
It’s a solid orange mass for each chromosome indicating that the child matches the parent at every location.
How can this be if the child only inherits half of the parent’s DNA?
Remember – the parent has two chromosomes that mix to give the child one chromosome. When comparing the child to the parent, the child’s single chromosome inherited from the parent matches one of the parent’s two chromosomes at every address location – so it shows as a complete match to the parent even though the child is only matching one of the parent’s two of chromosome locations. This isn’t a bug and it’s just how chromosome browsers work. In other words, the “other ” chromosome that your parents carry is the one you don’t match.
The diagram below shows the mother’s two copies of chromosome 1 she inherited from her father and mother and which section she gave to her child.
You can see that the mother’s father’s chromosome is blue in this illustration, and the mother’s mother’s chromosome is pink. The crossover points in the child are between part B and C, and between part C and D. You can clearly see that the child, when compared to the mother, does in fact match the mother in all locations, or parts, 3 blue and 1 pink, even though the source of the matching DNA is from two different parents.
This example shows the child compared to both parents, so you can see that the child does in fact match both parents on every single location.
This is exactly why two different matches may match us on the same location, but may not match each other because they are from different sides of our family – one from Mom’s side and one from Dad’s.
You can read more about this in the article, One Chromosome, Two Sides, No Zipper – ICW and the Matrix.
The only way to tell which “sides” or pieces of the parent’s DNA that the child inherited is to compare to other people who descend from the same line as one of the parents. In essence, you can compare the child to the grandparents to identify the locations that the child received from each of the 4 grandparents – and by genetic subtraction, which segments were NOT inherited from each grandparent as well, if one grandparent happens to be missing.
In our Parental Chromosome pink and blue diagram illustration above, the child did NOT inherit the pink parts A, B and D, and did not inherit the blue part C – but did inherit something from the parent at every single location. They also didn’t inherit an equal amount of their grandparents pink and blue DNA. If they inherited the pink part, then they didn’t inherit the blue part, and vice versa for that particular location.
The parent to child chromosome browser view also shows us that the very tip ends of the chromosomes are not included in the matching reports – because we know that the child MUST match the parent on one of their two chromosomes, end to end. The download or chart view provides us with the exact locations.
This brings us to the question of whether crossovers occur equally between males and female children. We already know that the X chromosome has a distinctive inheritance pattern – meaning that males only inherit an X from their mothers. A father and son will NEVER match on the X chromosome. You can read more about X chromosome inheritance patterns in the article, X Marks the Spot.
Crossovers Differ Between Males and Females
In the paper Genetic Analysis of Variation in Human Meiotic Recombination by Chowdhury, et al, we learn that males and females experience a different average number of crossovers.
The authors say the following:
The number of recombination events per meiosis varies extensively among individuals. This recombination phenotype differs between female and male, and also among individuals of each gender.
Notably, we found different sequence variants associated with female and male recombination phenotypes, suggesting that they are regulated by different genes.
Meiotic recombination is essential for the formation of human gametes and is a key process that generates genetic diversity. Given its importance, we would expect the number and location of exchanges to be tightly regulated. However, studies show significant gender and inter-individual variation in genome-wide recombination rates. The genetic basis for this variation is poorly understood.
The Chowdhury paper provides the following graphs. These graphs show the average number of recombinations, or crossovers, per meiosis for each of two different studies, the AGRE and the FHS study, discussed in the paper.
The bottom line of this paper, for genetic genealogists, is that males average about 27 crossovers per child and females average about 42, with the AGRE study families reporting 41.1 and the FHS study families reporting 42.8.
I have been collaborating with statistician, Philip Gammon, and he points out the following:
Male, 22 chromosomes plus the average of 27 crossovers = an average of 49 segments of his parent’s DNA that he will pass on to his children. Roughly half will be from each of his parents. Not exactly half. If there are an odd number of crossovers on a chromosome it will contain an even number of segments and half will be from each parent. But if there are an even number of crossovers (0, 2, 4, 6 etc.) there will be an odd number of segments on the chromosome, one more from one parent than the other.
The average size of segments will be approximately:
- Males, 22 + 27 = 49 segments at an average size of 3400 / 49 = 69 cM
- Females, 22 + 42 = 64 segments at an average size of 3400 / 64 = 53 cM
This means that cumulatively, over time, in a line of entirely females, versus a line of entirely males, you’re going to see bigger chunks of DNA preserved (and lost) in males versus females, because the DNA divides fewer times. Bigger chunks of DNA mean better matching more generations back in time. When males do have a match, it would be likely to be on a larger segment.
The article, First Cousin Match Simulations speaks to this as well.
What does this mean, practically speaking, to genetic genealogists?
Few lines actually descend from all males or all females. Most of our connections to distant ancestors are through mixtures of male and female ancestors, so this variation in crossover rates really doesn’t affect us much – at least not on the average.
It’s difficult to discern why we match some cousins and we don’t match others. In some cases, rather than random recombination being a factor, the actual crossover rate may be at play. However, since we only know who we do match, and not who tested and we don’t match, it’s difficult to even speculate as to how recombination affected or affects our matches. And truthfully, for the application of genetic genealogy, we really don’t care – we (generally) only care who we do match – unless we don’t match anyone (or a second cousin or closer) in a particular line, especially a relatively close line – and that’s a horse of an entirely different color.
To me, the burning question to be answered, which still has not been unraveled, is why a difference in recombination rates exists between males and females. What processes are in play here that we don’t understand? What else might this not-yet-understood phenomenon affect?
Until we figure those things out, I note whether or not my match occurred through primarily men or women, and simply add that information into the other data that I use to determine match quality and possible distance. In other words, information that informs me as to how close and reasonable a match is likely to be includes the following information:
- Total amount of shared DNA
- Largest segment size
- Number of matching segments
- Number of SNPs in matching segment
- Shared matches
- X chromosome
- mtDNA or Y DNA match
- Trees – presence, absence, accuracy, depth and completeness
- Primarily male or female individuals in path to common ancestor
- Who else they match, particularly known close relatives
- Does triangulation occur
It would be very interesting to see how the instances of matches to a certain specific cousin level – say 3rd cousins (for example), fare differently in terms of the average amount of shared DNA, the largest segment size and the number of segments in people descended from entirely female and entirely male lines. Blaine Bettinger, are you listening? This would be a wonderful study for the Shared cM Project which measures actual data.
Isn’t the science of genetics absolutely fascinating???!!!
I receive a small contribution when you click on some of the links to vendors in my articles. This does NOT increase the price you pay but helps me to keep the lights on and this informational blog free for everyone. Please click on the links in the articles or to the vendors below if you are purchasing products or DNA testing.
Thank you so much.
DNA Purchases and Free Transfers
- Family Tree DNA
- MyHeritage DNA only
- MyHeritage DNA plus Health
- MyHeritage FREE DNA file upload
- 23andMe Ancestry
- 23andMe Ancestry Plus Health
- Legacy Tree Genealogists for genealogy research
Dear Roberta, I am a devoted reader and thank you for all of your wonderful work. I just wondered if you had seen Newton Morton’s paper. It is at http://www.pnas.org/content/88/17/7474.full.pdf, search 88/17/7474. Females have more cMs on every arm of every chromosome unless there is none. Regards, Shedrick Moore, amateur.
I have not seen this paper. I wonder if this paper has been superseded with information from the Human Genome project. After I get back from Houston, I’ll focus on this more directly. Thank you.
That paper is from 1991 but it just goes to show that it has long been known that female parents tend to produce more crossovers on average than male parents. The estimates in the Morton paper (28.09 for males and 47.82 for females) weren’t too far from the more recently calculated values such as those in the Chowdhury paper from 2009 (around 27 for males and 42 for females). Not bad considering that in 1991 they didn’t have the benefit of modern DNA sequencing techniques!
You’ve been reading my mind … I was recently hunting around online for a good concise explanation of these two terms – and now you’ve provided it! Many thanks once again 🙂
does age of the parent affect the number of crossovers ?
I don’t know if that has ever been studied. I don’t have an answer. I will see what I can find.
Age would not effect the maternal side, as the oocytes (Ova/egg) are formed during fetal development. Meaning crossover have already occurred.
“During fetal life, there are about 6 million to 7 million eggs. From this time, no new eggs are produced. At birth, there are approximately 1 million eggs; and by the time of puberty, only about 300,000 remain. Of these, only 300 to 400 will be ovulated during a woman’s reproductive lifetime.”
For men it may be another story, as sperm are constantly being produced.
True, and I just read something today about mutations occurring in older fathers.
I have asked Philip what he knows. He may comment on the article.
I am not positive about my above statement about when the crossovers happen in the oocytes, It may be just before the mature Egg is formed, so age could also have an effect….
Hi prairielad, I’m not a biologist but I think that your second posting is correct and that the final step in the process occurs just before the egg is formed. In females “oocytes remain arrested in a late stage of prophase (dictyotene) for many years, if not decades”. The number of crossovers does increase with maternal age. It’s only a very gradual increase, an average of an additional 0.05 crossovers per year until around the age of 39, when the average number of crossovers increases by about 3 per year. No age effect is observed for males.
With mutations it is the opposite. There is no observed age effect in females but a steady increase in the number of mutations with age for males due to the higher number of cell divisions that occur over time.
“It would be very interesting to see how the instances of matches to a certain specific cousin level – say 3rd cousins (for example), fare differently in terms of the average amount of shared DNA, the largest segment size and the number of segments in people descended from entirely female and entirely male lines. Blaine Bettinger, are you listening? This would be a wonderful study for the Shared cM Project which measures actual data.”
Yes!! Yes!!! Please!
I have wondered whether this explains the seeming (anecdotal) difference between Y matches and mt matches. In my limited experience, there are fewer mt matches who are also autosomal matches than Y matches who are also autosomal matches. I have no hard proof of this, given my small sample size, but would like to see some.
You do not say how to find crossover points when grandparents have not been tested. My grandparents have passed. However, both sisters, both parents, 1 maternal aunt, and 3 paternal uncles have been tested. How, then, can I determine the crossover points? Thanks
Do you think the Dnagen.net function will be back for FTDNA Family Finder?
Enjoy your blog very much. Keep up the good work.
It’s being worked on collaboratively.
I just received my uncle’s result from FTDNA, and I am kinda lost as there is a huge discrepancy between his information and my mother’s in spite of them being full siblings. For instance, He is 27%Iberia whereas mom is 7%; 18% British isles, mother is 0%; 0% Central Europe my mother is 38%.
I know that differences appear, but aren’t those results too far distant for full siblings? I have seen ppl posting their results on Youtube, and there’s just a <1 to 2 % differences between siblings. So that came as a shocker for me today. Could you help me understand, please, Mrs Estes.
You should never utilize ethnicity as an indicator of parentage.
Roberta, I was very surprised to see that your initial example showed that 6 of 22 autosomes were passed entire from one of the two grandparents. I realize it’s random, but do you have any data on how often a chromosome has 0 crossovers (meaning nothing from one chromosome in a pair is passed on)? I have no data on the topic, but 6 times is certainly higher than I would have expected/guessed.
No, I don’t have any hard data, but I see it a lot.
Kevin, I have some data. You can contact me at firstname.lastname@example.org
If you want to publish something specific to this topic, let me know.
Pingback: First Steps When Your DNA Results are Ready – Sticking Your Toe in the Genealogy Water | DNAeXplained – Genetic Genealogy