Genetics
Crossing Over
Homologous chromosomes physically swap matching DNA segments during meiosis I — ~30 crossovers per human meiosis
Crossing over is the reciprocal exchange of DNA between homologous chromosomes during meiosis I, swapping matching segments at structures called chiasmata. It begins with a programmed SPO11 double-strand break, is repaired by homologous recombination, and reshuffles parental alleles into new combinations. Humans average 1–2 crossovers per chromosome arm (~30 per meiosis); failure to recombine is a leading cause of aneuploidy like Down syndrome.
- WhenProphase I (leptotene–pachytene)
- Initiated bySPO11 double-strand break
- Breaks per meiosis~200–300 (humans)
- Crossovers per meiosis~30 (≥1 per bivalent)
- Resolved atHolliday junction → chiasma
- Failure causeNondisjunction / aneuploidy
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
What crossing over is, intuitively
Picture the two copies of a chromosome you inherited — one from your mother, one from your father. They carry the same genes in the same order, but with different alleles: maybe your mother's copy has the eye-color variant A and the height variant B, while your father's has a and b. If those chromosomes were simply handed down intact, every gamete you make would pass on either the whole maternal block or the whole paternal block. Crossing over breaks that rule. During meiosis, the two homologs line up alongside each other, are physically cut, and trade matching segments — so a single chromosome that leaves your body in a sperm or egg can carry your mother's A next to your father's b. That recombined chromosome is something neither of your parents ever had.
The "matching" part is critical. Crossing over is not a random splice; it is template-guided repair. The cut chromosome searches the homolog for the same sequence and exchanges DNA precisely at that point, so no genes are lost or duplicated — the gene order is preserved, only the allele combinations change. The visible result, when you stain a cell in mid-meiosis and look down a microscope, is an X-shaped crossing point between two of the four chromatids: a chiasma. That little cross is the physical signature of an exchange that just rewrote inheritance.
The mechanism, step by step
Crossing over is a controlled act of DNA self-injury followed by templated repair. The molecular sequence in a human (or yeast, where most of this was worked out) meiocyte runs like this:
- Programmed break (leptotene). The conserved enzyme SPO11, a relative of archaeal topoisomerase VI working with the partner TOPOVIBL, cleaves both strands of the DNA. Unlike accidental breaks, these are deliberate and numerous — about 200–300 double-strand breaks per human meiocyte. SPO11 stays covalently bonded to the cut 5′ ends.
- End processing. The MRN complex (MRE11–RAD50–NBS1) together with CtIP nicks the DNA and releases SPO11 still attached to a short oligonucleotide. Exonucleases then chew back the 5′ strands (resection), exposing single-stranded 3′ overhangs hundreds of nucleotides long.
- Homology search and strand invasion. The recombinases RAD51 and the meiosis-specific DMC1 coat the 3′ tails to form nucleoprotein filaments. These filaments probe the homologous chromosome for matching sequence and, when they find it, invade the duplex — displacing one strand and base-pairing with the other to form a displacement loop (D-loop).
- Synapsis and stabilization (zygotene–pachytene). Meanwhile the homologs zip together along their length via the synaptonemal complex, a protein ladder roughly 100 nm wide with transverse filaments (SYCP1 in mammals) bridging two lateral elements. This scaffold holds the partners in register while recombination matures.
- DNA synthesis and junction formation. A DNA polymerase extends the invading 3′ end using the homolog as template. Capture of the second break end and further synthesis and ligation produce a double Holliday junction — a four-stranded, fully base-paired intermediate linking the two chromatids.
- Resolution. The MutSγ complex (MSH4–MSH5) marks intermediates destined to become crossovers, and the MutLγ endonuclease (MLH1–MLH3) cuts the junction in the orientation that exchanges the flanking arms. The result is a reciprocal crossover. Intermediates not selected for this fate are dissolved or resolved as non-crossovers (gene conversions) that copy a small patch without swapping arms.
- Chiasma and segregation (diplotene–anaphase I). The synaptonemal complex disassembles, but each crossover plus distal sister-chromatid cohesion leaves a chiasma physically holding the homologs together. That tension lets the spindle correctly bi-orient the bivalent; cohesion release at anaphase I then pulls the recombined homologs to opposite poles.
The players and conditions
- SPO11 / TOPOVIBL. The "molecular scissors" that initiate every meiotic crossover by making the deliberate double-strand break. Deleting SPO11 abolishes recombination and causes sterility.
- PRDM9. In humans, mice, and most mammals, this rapidly evolving zinc-finger protein chooses where breaks happen. It binds specific DNA motifs and deposits the histone marks H3K4me3 and H3K36me3 that recruit the SPO11 machinery, defining tens of thousands of narrow recombination hotspots. Mismatches in PRDM9 binding between subspecies are a known cause of hybrid sterility.
- RAD51 and DMC1. The recombinases that perform the homology search and strand invasion. DMC1 is meiosis-specific and biases repair toward the homolog rather than the sister chromatid — essential for inter-homolog crossing over.
- Synaptonemal complex (SYCP1/SYCP2/SYCP3). The structural scaffold that aligns homologs and helps regulate crossover number and spacing.
- MutSγ (MSH4–MSH5) and MutLγ (MLH1–MLH3). The crossover-designation and resolution machinery; MLH1 foci are the standard cytological count of crossovers.
- Conditions. Crossing over requires a homologous partner (it does not occur in normal mitosis), occurs only in germ-line cells entering meiosis, and is subject to two regulatory rules: the obligate crossover (at least one per bivalent) and crossover interference (one crossover suppresses others nearby, spacing them out).
Crossing over vs independent assortment vs gene conversion
| Property | Crossing over | Independent assortment | Gene conversion (non-crossover) |
|---|---|---|---|
| What shuffles | Alleles within one chromosome | Whole chromosomes between pairs | A short DNA tract (no arm swap) |
| When | Prophase I | Metaphase I | Prophase I |
| Mechanism | Reciprocal DNA exchange at a chiasma | Random spindle orientation of bivalents | Non-reciprocal copy from homolog |
| Reciprocal? | Yes — both chromatids change | N/A | No — one chromatid copies the other |
| Combinations (human) | Effectively unlimited along a chromosome | 223 ≈ 8.4 million | Local, allele-level only |
| Visible sign | Chiasma (X-shaped junction) | Bivalent alignment at the plate | None cytologically |
| Fraction of DSBs | ~10% of breaks (humans) | N/A | ~90% of breaks |
| Breaks linkage? | Yes — separates linked loci | Yes — across chromosomes | Rarely (very short tract) |
Real numbers and scales
- Double-strand breaks: ~200–300 are deliberately made per human meiocyte by SPO11.
- Crossovers: ~30 on average per human meiosis (roughly 1–2 per chromosome arm; at least 1 per bivalent). Counts differ by sex — human female meiosis averages ~42 crossovers, male ~27 — a difference reflected in the longer female genetic map (~4,460 cM vs ~2,590 cM).
- Hotspots: 25,000–50,000 PRDM9-defined hotspots, each a few kilobases wide, account for most crossovers; the genome between them is largely cold.
- Map vs physical distance: the human genome is ~3,200 Mb of DNA but ~3,600 cM of sex-averaged genetic map, so 1 cM ≈ 0.9 Mb on average — but recombination rate varies more than 100-fold along a chromosome.
- Centimorgan definition: 1 cM = a 1% recombination frequency between two loci per meiosis. Loci more than ~50 cM apart recombine ~50% of the time and behave as if unlinked.
- Synaptonemal complex: ~100 nm between lateral elements; assembles over hours in zygotene and is fully formed in pachytene.
- Gene-conversion tracts: typically a few hundred base pairs to ~1–2 kb of copied sequence.
- Timescale: in human males meiosis completes in ~64 days; in females, oocytes enter prophase I before birth and can remain arrested at the chiasma stage for up to ~50 years before ovulation.
Where it shows up — organisms, disease, technology
- You and every sexually reproducing eukaryote. From yeast to maize to humans, crossing over is the engine of recombinant gametes. It is the reason no two siblings (except identical twins) are genetically the same.
- Down syndrome and aneuploidy. Chromosome 21 bivalents that fail to make a crossover — or recombine too near the centromere or telomere — are strongly linked to maternal nondisjunction and trisomy 21. The same recombination failures drive trisomy 18, trisomy 13, and the majority of early miscarriages, and the risk rises with maternal age as chiasma-stabilizing cohesion decays.
- Genetic mapping and gene hunting. Recombination frequency is the ruler of classical genetics. Sturtevant's 1913 Drosophila map, human linkage studies, and modern fine-mapping of disease loci all read out crossover frequencies as centimorgans.
- Crop and livestock breeding. Breeders depend on crossing over to combine favorable alleles; "linkage drag," where a desired gene stays stuck next to a deleterious neighbor, is a direct consequence of low local recombination, and tools to boost meiotic crossovers (for example knocking out the anti-crossover helicase FANCM) are an active breeding target.
- Disease from misfired exchange. Crossing over between non-allelic repeated sequences (non-allelic homologous recombination) produces deletions and duplications such as the 1.4 Mb duplication causing Charcot-Marie-Tooth disease type 1A and its reciprocal deletion causing hereditary neuropathy with pressure palsies.
- Cancer biology connection. The same homologous-recombination machinery (RAD51, the BRCA1/BRCA2 mediators) repairs breaks in somatic cells; its loss underlies BRCA-mutant cancers and their sensitivity to PARP inhibitors.
Common misconceptions
- "Crossing over happens between sister chromatids." Productive meiotic crossovers occur between non-sister chromatids of homologous chromosomes — that is what reshuffles different parental alleles. Sister chromatids are identical, so exchanging them changes nothing. The DMC1 recombinase specifically biases repair toward the homolog.
- "It happens in mitosis too." Routine crossing over is a meiotic program. Mitotic cells do use homologous recombination to repair breaks, but they preferentially use the identical sister chromatid and crossovers are actively suppressed there to avoid loss of heterozygosity.
- "More breaks means more crossovers." Only ~10% of the 200–300 breaks become crossovers; the rest are repaired as non-crossover gene conversions. Crossover number is tightly regulated by the obligate-crossover rule and crossover interference, not by break count alone.
- "Genes lost or gained during the swap." Because the exchange is homology-guided, gene order and dosage are preserved — only allele combinations change. Loss or gain happens only in the pathological case of misaligned, non-allelic recombination.
- "Crossing over and independent assortment are the same shuffling." They are complementary but distinct. Independent assortment randomizes which whole chromosome of each pair goes to a gamete; crossing over rearranges alleles within a chromosome. Both are needed to explain the full diversity of gametes.
- "A chiasma is just a tangle." A chiasma is the load-bearing physical link that, with cohesion, generates the tension the spindle needs. Lose it and the homologs can co-migrate — the classic route to nondisjunction.
Frequently asked questions
When during meiosis does crossing over happen?
Crossing over occurs in prophase I of meiosis, specifically across the substages leptotene through pachytene. In leptotene, the topoisomerase-like enzyme SPO11 introduces programmed double-strand breaks across the genome — about 200–300 in a human meiocyte. In zygotene, homologous chromosomes pair and the synaptonemal complex, a zipper-like protein scaffold about 100 nm wide, begins to assemble between them. By pachytene the homologs are fully synapsed and the recombination intermediates mature; a subset are resolved as crossovers. In diplotene the synaptonemal complex disassembles and the homologs pull apart but stay tethered at chiasmata — the cytological sign of a completed crossover — which hold the bivalent together until the homologs separate at anaphase I. So although the DNA breaks are made early, the visible exchange points persist for weeks to decades in human oocytes arrested in this stage.
How is a crossover different from a chiasma?
A crossover is the molecular event — the reciprocal exchange of DNA between two non-sister chromatids, created when a double-strand break is repaired using the homologous chromosome as a template and resolved in the crossover orientation. A chiasma is the physical, microscope-visible cross-shaped junction where the exchanged chromatids remain connected. Every chiasma reflects an underlying crossover, but not every double-strand break becomes a crossover: in humans only about 10 percent of the 200–300 initial breaks mature into crossovers, while the rest are repaired as non-crossovers (gene conversions) that copy a small patch of sequence without exchanging flanking arms. Chiasmata are mechanically essential — together with sister-chromatid cohesion they create the tension that lets the spindle correctly orient and separate homologs at anaphase I.
What molecules actually perform crossing over?
The exchange is built by the homologous recombination machinery. SPO11, aided by partners like TOPOVIBL, cuts the DNA and stays covalently attached to the 5' ends until the MRN complex (MRE11-RAD50-NBS1) plus CtIP nick and free it. Exonucleases then resect the 5' strands, leaving 3' single-stranded overhangs that are bound by the recombinases RAD51 and the meiosis-specific DMC1. These recombinase filaments search the homologous chromosome for matching sequence and catalyze strand invasion, forming a displacement loop (D-loop). DNA polymerase extends the invading strand; capture of the second end and ligation build a double Holliday junction. The MutLgamma endonuclease (MLH1-MLH3), guided by the MSH4-MSH5 (MutSgamma) complex, resolves most mammalian crossovers, leaving MLH1 foci that geneticists count to measure crossover number — typically about 50 per human pachytene spermatocyte (scoring the whole bivalent) and roughly 30 per meiosis on average when sex-averaged.
Why does crossing over matter for evolution?
Crossing over breaks the physical linkage between alleles on the same chromosome, so beneficial and harmful mutations that arose in different individuals can be brought together (or separated) in a single genome. Without recombination, a chromosome behaves as one giant non-dividing block: a good gene is stuck forever next to whatever bad genes happen to share its DNA, and natural selection can only act on the whole block at once. Recombination lets selection 'see' each locus more independently, accelerates adaptation, and prevents the irreversible accumulation of deleterious mutations known as Muller's ratchet. It is also the reason siblings differ: combined with independent assortment of 23 chromosome pairs (2^23 ≈ 8.4 million combinations), crossing over makes the number of genetically distinct gametes a human can produce effectively unlimited. The trade-off is that recombination can also break up favorable combinations and occasionally causes harmful rearrangements when it happens between misaligned repeats.
How does crossover frequency relate to genetic maps?
The probability that two loci are separated by a crossover rises with the physical distance between them, which is the basis of genetic mapping. The unit of map distance is the centimorgan (cM): two loci are 1 cM apart if a crossover occurs between them in 1 percent of meioses. Thomas Hunt Morgan and his student Alfred Sturtevant exploited this in 1913, when Sturtevant — as an undergraduate — used recombination frequencies among Drosophila X-linked genes to draw the first genetic map, placing genes in linear order. The human genome is about 3,200 Mb of DNA but roughly 3,600 cM of genetic map, so on average about 1 cM corresponds to ~0.9 Mb — though the conversion varies enormously, since crossovers cluster at ~25,000–50,000 recombination hotspots a few kilobases wide. In humans these hotspots are positioned by the zinc-finger protein PRDM9, which writes a histone mark (H3K4me3 and H3K36me3) that recruits SPO11.
What happens if crossing over fails?
Each pair of homologs needs at least one crossover — the 'obligate crossover' — to segregate correctly. A bivalent with no chiasma (an achiasmate pair) has nothing holding the homologs together, so they can drift apart and end up moving to the same pole. The result is nondisjunction: gametes with an extra or missing chromosome. This is the single largest cause of human aneuploidy. Maternal chromosome 21 bivalents that fail to recombine, or that recombine too close to the centromere or telomere, are strongly associated with trisomy 21 (Down syndrome); similar recombination errors underlie trisomy 18, trisomy 13, and most miscarriages. The risk climbs steeply with maternal age because oocytes are arrested in prophase I for decades, during which the cohesion that maintains chiasma positioning gradually deteriorates. Conversely, too much crossing over near a centromere, or recombination between non-allelic repeated sequences, can generate deletions, duplications, and translocations that cause disorders such as Charcot-Marie-Tooth disease type 1A.