Horizontal Gene Transfer: How Microbes Swap DNA

Most animals get their genes from two parents and nowhere else. Bacteria do not follow that rule. Horizontal gene transfer (HGT) is the movement of DNA between organisms that are not parent and offspring, and microbes manage it in at least three distinct ways: by absorbing loose DNA from their surroundings, by linking up and copying DNA across a tiny bridge, and by being infected with viruses that carry DNA from a previous victim. This article walks through each route, the experiments that found them, and what the process did to the tree of life.

Two ways to inherit a gene

Vertical inheritance is the familiar kind. A cell divides, and both daughters get a copy of the parent's chromosome, errors and all. Every gene you carry arrived this way, from a lineage of ancestors stretching back through time. If you could trace any gene far enough, it would lead to a single line of descent.

Horizontal inheritance skips the family line. A gene is copied out of one cell and into another cell that is not its descendant, sometimes into a cell of a different species, genus, or even domain of life. The recipient does not need to wait for a lucky mutation. It receives a working gene that has already been tested by natural selection somewhere else.

A useful comparison is the difference between writing a program from scratch and installing an app. Mutation is the slow route, where small edits to existing code accumulate over thousands of generations until a new function appears. HGT is closer to sideloading a finished program written by someone else. The code may not run perfectly in its new home, and it can crash the system, but when it works the gain is immediate.

The same comparison explains why the process matters so much to biologists. In 2000, Howard Ochman, Jeffrey Lawrence, and Eduardo Groisman argued in a review in Nature that bacteria gain much of their diversity by acquiring sequences from distantly related organisms, in contrast to eukaryotes, where change mostly comes from modifying genes that were already there. That review also described HGT as producing extremely dynamic genomes, and noted that lateral transfers have changed the ecological and disease-causing character of bacterial species (Ochman, Lawrence, and Groisman, Nature, 2000).

Key idea: HGT is not a rare accident in microbes. In the E. coli lineage, a 1998 genome analysis estimated that about 18% of the current chromosome was acquired from other organisms over roughly 100 million years.

How much of a genome can be borrowed

The most cited measurement comes from a 1998 paper by Jeffrey Lawrence and Howard Ochman, "Molecular archaeology of the Escherichia coli genome," published in the Proceedings of the National Academy of Sciences. They looked for stretches of DNA whose composition did not match the rest of the chromosome, a sign that the stretch came from somewhere else. They concluded that 755 of the 4,288 protein-coding genes (open reading frames) in the E. coli K-12 genome had been introduced in at least 234 separate transfer events since E. coli diverged from the Salmonella lineage, roughly 100 million years ago. Most of the acquired genes were later deleted, and the roughly 18% that persisted gave E. coli access to ecological niches it could not otherwise have occupied (Lawrence and Ochman, PNAS, 1998).

Two details in that result are easy to miss. First, the 18% is what survived. Most acquired genes were later lost, so the number of transfers over time was higher than the number of foreign genes visible today. Second, this method only detects transfers whose DNA still looks foreign. An ancient transfer that has had time to blend into the host's genetic style would be missed, so the true figure is probably higher for old events and the method is best read as a lower bound.

Transferred genes are also not spread evenly. A review of the field notes that horizontally transferred genes are typically concentrated in about 1% of a bacterial chromosome (Wikipedia summary of the primary literature on HGT). Those hotspots work like a junk drawer for new equipment. They hold the genes for toxins, drug resistance, and unusual metabolism, while the core machinery for copying DNA and making proteins sits elsewhere and changes much more slowly.

That split gives microbiologists a practical way to describe a species. The "core genome" is the set of genes shared by every strain. The "accessory genome" is the set that varies, and much of it was acquired horizontally. Two E. coli strains can share a core and still differ enough in their accessory genes that one lives harmlessly in the gut and the other causes severe disease.

Transformation: picking up loose DNA

Transformation is the oldest route to be identified, and it is conceptually the simplest. A bacterium takes up naked DNA from its surroundings and, if the sequence is similar enough to something in its own chromosome, swaps it in. The DNA comes from cells that have died and broken open, or in some species from cells that actively release it.

Frederick Griffith saw the effect first without knowing what he was looking at. In 1928 he was studying pneumonia-causing pneumococcus bacteria in mice. A virulent smooth strain (type III-S) killed mice, while a nonvirulent rough strain (type II-R) did not. When he injected heat-killed III-S bacteria together with live II-R bacteria, the mice died, and live III-S bacteria could be recovered from their blood. Griffith attributed this to a "transforming principle" released from the dead cells (overview of Griffith's experiment). The full story of that experiment, and the way it prepared the ground for the discovery that DNA is the genetic material, is told in the Griffith transformation experiment article. The principle was later identified as DNA in the Avery-MacLeod-McCarty experiment, and the result was reinforced by the Hershey-Chase experiment.

The sequence of discovery is worth noticing. HGT was the first thing observed, and it was observed before anyone knew what a gene was made of. The experiment that first suggested bacteria could transfer genetic information is the same experiment that eventually led to the identification of DNA. In other words, the sideways movement of genes is not an obscure add-on to molecular biology. It sits at the foundation of it.

What "competence" means

Not every bacterium can pick up DNA at will. A cell that can is called competent, and competence is usually a regulated state. In Bacillus subtilis, for example, becoming competent requires the expression of about 40 genes, which build a machine that captures DNA outside the cell, feeds one strand through the membrane, and protects it on the inside. Natural transformation has been documented in at least 67 prokaryotic species (Wikipedia summary of the primary literature).

Think of competence as a cell opening a mail slot. Most of the time the slot is shut. When conditions change, for example when the population is crowded or under stress, the cell opens the slot and starts reading whatever lands on the mat. If a letter contains a useful gene and the sequence at each end matches the cell's own chromosome, the cell can paste the new text over the old one.

Researchers also use transformation on purpose. In a laboratory, scientists make cells artificially competent with heat or electricity so that they will accept a plasmid carrying a gene of interest. The natural mechanism and the lab technique are the same trick at different scales.

Conjugation: a bridge between two cells

If transformation is mail, conjugation is a handshake. Two bacteria connect physically, and a copy of DNA moves from the donor to the recipient through that connection.

Joshua Lederberg and Edward Tatum reported the effect in a short letter in Nature titled "Gene Recombination in Escherichia coli" on October 19, 1946 (Nature 158, 558). They worked with two mutant strains of E. coli K-12, each unable to make certain nutrients it needed to grow, so each strain starved on a minimal medium. When they mixed the two strains and plated the mixture on that minimal medium, a few colonies grew. Colonies grew only from the mixed cultures, never from either strain alone. Those colonies had to carry the working versions of genes from both parents, which meant that genetic material had recombined between them. Lederberg shared the 1958 Nobel Prize in Physiology or Medicine with Tatum and George Beadle "for his discoveries concerning genetic recombination and the organization of the genetic material of bacteria."

The design of the experiment is an example of how to measure a rare event. A billion cells that cannot grow is a clean background. If even one in ten million manages to grow, the colonies are impossible to miss. The same logic, in which selection removes everything except the event of interest, is at work in many classic genetics experiments, including the Luria-Delbrück fluctuation test.

The F plasmid and the pilus

What is the bridge made of? In the classic E. coli system, the donor carries a plasmid called the F factor (fertility factor), a circle of DNA of about 100 kilobases that replicates independently of the main chromosome. A region of about 33 kilobases, containing around 40 genes, directs construction of the F pilus, a thin filament that reaches out to a recipient cell and pulls the two together. The pili are described as extremely resistant to mechanical and thermochemical stress (overview of bacterial conjugation).

Once the cells are in contact, the plasmid is nicked at a specific site, and one strand is fed across to the recipient while the donor makes a replacement strand. The recipient does the same with its half. At the end, both cells have the plasmid, and the recipient has become a donor. In practical terms, a plasmid is a piece of software that copies itself between machines and installs on arrival.

Sometimes the plasmid is not floating freely but has integrated into the main chromosome. Such strains are called Hfr, for "high frequency of recombination," because they transfer chromosomal genes far more efficiently. In common laboratory strains, moving the whole chromosome takes about 100 minutes. Researchers exploited that slow, steady pace. By separating mating cells with a kitchen blender at different times and checking which genes had crossed, they could map the order of genes along the chromosome, treating the clock as a ruler.

Conjugation across species

Conjugation does not always stay inside a species. Some plasmids can transfer between distantly related bacteria, and the transfer can move DNA from bacteria into other kinds of cells. The best-known example is the plant pathogen Agrobacterium, which uses a conjugation-like system to deliver a piece of DNA (T-DNA) into plant cells. That delivery system has had large consequences, both in nature and in biotechnology. It is the reason that, as discussed below, sweet potatoes carry bacterial genes.

Transduction: viruses as couriers

The third route recruits a virus. Bacteriophages ("phages") are viruses that infect bacteria. A phage injects its genome into a cell and takes over the cell's machinery to build new phage particles. Sometimes, by accident, the phage packages a fragment of the host's DNA into a new particle instead of its own genome. When that particle infects another cell, it delivers the previous victim's DNA.

Norton Zinder and Joshua Lederberg discovered this process in Salmonella in 1952 at the University of Wisconsin-Madison, and named it transduction (overview of transduction). The shipment is a mistake, not a design, which is why it is a poor courier by human standards. In generalized transduction, any part of the host chromosome can end up in a phage head, so the cargo is random. Generalized transduction is rare, on the order of one phage particle in 11,000.

Specialized transduction is more selective. A phage that has integrated its genome into the host chromosome (a prophage) may later cut itself out imprecisely, taking a few neighboring host genes with it. The textbook example is phage lambda in E. coli, which carries the genes next to its insertion site.

Imagine a delivery truck that occasionally loads boxes from the wrong warehouse. Each mistake is rare, but a city with trillions of trucks and a constant stream of deliveries will see it happen all the time. Phages are thought to be the most abundant biological entities on Earth, so even at low error rates, a large amount of DNA moves around the planet this way.

When a phage carries a poison

Transduction and its relatives matter most when the DNA that moves is dangerous. In 1996, Matthew Waldor and John Mekalanos showed that the genes for cholera toxin are carried by a filamentous phage called CTXφ, which is related to the coliphage M13. The phage can integrate into the chromosome of Vibrio cholerae or replicate as a plasmid, and it uses a surface structure on the bacterium (the toxin-coregulated pilus) as its receptor. A harmless strain of V. cholerae infected by CTXφ becomes a toxin producer. The authors reported that infection occurred more efficiently in the mouse intestine than in the laboratory, and concluded that horizontal transfer may underlie the emergence of toxigenic V. cholerae (Waldor and Mekalanos, Science, 1996).

That finding changed how people think about bacterial disease. Cholera is not just a bacterium. It is a bacterium plus a virus that carries the instructions to poison the gut. An outbreak strain can be assembled in the environment out of parts that were harmless on their own, in a process the same paper described as lysogenic conversion. (Lysogeny is the state in which a phage genome sits quietly inside a host cell.) Louis Pasteur and Robert Koch built the idea that each disease has a germ behind it (germ theory), and HGT adds a twist: the germ itself can be a moving target, acquiring and losing the genes that make it dangerous.

The three routes side by side

The table below compares the three main mechanisms.

Feature	Transformation	Conjugation	Transduction
What carries the DNA	Free DNA in the environment	Direct cell-to-cell contact (pilus)	A bacteriophage particle
Who decides	Recipient (competence)	Donor (plasmid-encoded machinery)	Neither (packaging accident)
Typical cargo	Fragments of chromosome	Plasmids, sometimes chromosome (Hfr)	Random host DNA or genes next to a prophage
First shown	Griffith, 1928	Lederberg and Tatum, 1946	Zinder and Lederberg, 1952
Needs live donor	No	Yes	Donor cell is usually killed by the phage
Typical distance between donor and recipient	Often close relatives	Often broad, depending on the plasmid	Limited by which cells the phage can infect

A fourth category deserves mention, even if it is less tidy. Mobile genetic elements, such as transposons ("jumping genes," the discovery that earned Barbara McClintock her Nobel Prize) and integrative elements, move DNA around inside a genome and can hitch a ride on any of the three routes above. A plasmid carrying a transposon that carries a resistance gene is a set of nested Russian dolls, and each doll can move on its own schedule.

The evidence that antibiotic resistance travels

Public health authorities watch HGT because it is how drug resistance moves between species. The Centers for Disease Control and Prevention (CDC) notes that resistant germs can transfer their defenses to germs that were never exposed to the drugs, and that resistance has been found in all regions of the world. The agency's 2019 report estimated more than 2.8 million antimicrobial-resistant infections each year in the United States, with more than 35,000 deaths (CDC).

The link between resistance and plasmids was first noticed in Japan. Infectious multiple drug resistance was recognized there in 1959, when dysentery bacteria (Shigella) turned out to carry resistance to several antibiotics at once, and the resistance could pass to other bacteria in the gut. The genes were not on the bacterial chromosome but on separate, self-replicating elements that became known as R factors. Tsutomu Watanabe's 1963 review in Bacteriological Reviews, "Infective Heredity of Multiple Drug Resistance in Bacteria," brought the topic to a wide audience. A later survey of Shigella in Japan found that most resistant strains carried R factors that could pass to other bacteria by cell-to-cell contact.

The practical consequence is that a single plasmid can make a bacterium resistant to several drugs in one step. A cell does not need to evolve resistance to each drug separately. It needs to receive one package. That is why treatments that work against a given bacterium can fail abruptly, and why resistance that arises in a harmless gut bacterium can reach a dangerous pathogen. The full picture of how that unfolds in hospitals, farms, and the environment is covered in the articles on how antibiotic resistance spreads and why antibiotics are running out. Here the point is narrower: resistance is a mobile trait, and the mechanisms in this article are the means of transport.

Beyond bacteria: animals, plants, and the cases that got retracted

HGT among bacteria is well established. HGT into complex organisms is rarer, harder to prove, and the source of a few embarrassing reversals. The cases below are ordered from solid to disputed.

Bdelloid rotifers. These microscopic freshwater animals, with nearly 500 described species and no known males, were examined in a 2008 Science study by Eugene Gladyshev, Matthew Meselson, and Irina Arkhipova. They found many genes that seemed to come from bacteria, fungi, and plants, clustered near chromosome ends together with mobile genetic elements. The gene-dense regions near the middle of the chromosomes appeared to be free of foreign genes. Some foreign genes were broken, while others were intact and active, and some contained working spliceosomal introns. One gene of apparent bacterial origin was expressed in E. coli and yielded an active enzyme (Gladyshev, Meselson, and Arkhipova, Science, 2008). The authors suggested a mechanism: bdelloids survive drying out, during which their DNA is fragmented and cell membranes are disrupted, and when they rehydrate they repair both, which may let them absorb DNA from their surroundings (ScienceDaily summary). The idea is a hypothesis, not a confirmed mechanism. One later estimate puts the bacterial-origin share of bdelloid genes at about 8%, which makes them the animal record holder.

Pea aphids. These insects carry genes of fungal origin. The red pigment torulene in aphids is the only carotenoid known to be made by an animal, and it depends on genes thought to have come from a fungus. The adzuki bean beetle, meanwhile, holds genetic material that it acquired from the bacterial endosymbiont Wolbachia.

Sweet potato. A 2015 study in PNAS by Tina Kyndt and colleagues found a stretch of T-DNA from Agrobacterium, the same system described under conjugation above, in all 291 cultivated sweet potato accessions they tested, but not in close wild relatives. They characterized two distinct T-DNA regions, called IbT-DNA1 and IbT-DNA2 (the second was found in 45 of 217 genotypes), and showed that genes in them are expressed in several tissues. The authors considered this an example of a naturally transgenic food crop (Kyndt et al., PNAS, 2015). The finding is a useful reminder that "genetically modified" has a natural precedent, though it says nothing about whether any particular engineered crop is safe.

Ants. A comparative analysis of ant genomes identified 497 horizontally transferred protein-coding genes in 85 species, some conserved across species for as long as 40 million years.

Humans. One proposal held that about 100 of roughly 20,000 human genes came from HGT, but other researchers have challenged that figure, and the cases remain contested.

Tardigrades. In 2015, Boothby and colleagues reported that roughly 17% of the genes in the tardigrade Hypsibius dujardini came from HGT, a figure that attracted wide attention. A second group, led by Georgios Koutsovoulos, resequenced and filtered the genome and concluded that most of those candidates were contaminants from bacteria sequenced along with the animal. Their estimate was that HGT accounts for at most 1% to 2% of tardigrade genes (Koutsovoulos et al., PNAS, 2016, 113:5053-5058).

The tardigrade story is a good lesson in how to read HGT claims. A genome assembled from a mixed sample can look as if an animal is stuffed with bacterial genes when the bacteria are simply in the tube. Credible claims need evidence that a foreign gene sits inside animal DNA, flanked by animal genes, and is present in independent samples.

What HGT did to the tree of life

For about a century after Darwin, biologists drew evolution as a tree. Species branch from common ancestors and never rejoin. Carl Woese's work on ribosomal RNA in the 1970s and 1980s gave the tree a molecular form and introduced the idea of a last universal common ancestor from which the three domains of life descend.

When genomes began to be sequenced in the 1990s, trees built from different genes often disagreed with each other. A gene for one function would group a bacterium with one set of relatives, and a gene for a different function would group it with another. In February 2000, W. Ford Doolittle summarized the problem in Scientific American in an article titled "Uprooting the Tree of Life," which argued that the family tree of life is more complicated than once believed and may not have had a single root (Doolittle, Scientific American, 2000). A 2015 review by Shannon Soucy, Jinling Huang, and Peter Gogarten, titled "Horizontal gene transfer: building the web of life," put the new picture in its title (Nature Reviews Genetics, 2015).

It would be wrong to conclude that the tree is dead. Critics point out that HGT does not invalidate phylogenetics. For animals and plants, vertical inheritance dominates, and the tree remains a good description. Even for microbes, the most reasonable summary is that a tree describes the core genes and a web describes the accessory ones. A rough analogy is a family surname. The surname passes down a clean line, even though the family's recipes, books, and tools are borrowed from neighbors. If you wanted to know who a person descends from, you would follow the surname. If you wanted to know how the household functions, you would look at the borrowed things.

Genes that moved inside a single cell

One of the most important cases of gene transfer did not involve two separate organisms. Mitochondria are descended from bacteria that took up residence inside an ancestral cell, an idea championed by Lynn Margulis. Over time, many genes that were once in the mitochondrial genome moved to the nucleus of the host, a process called endosymbiotic gene transfer. The story of how the symbiosis happened is told in how mitochondria became cell powerhouses. It is HGT at the scale of an entire organelle, and a reminder that the boundary between "your genes" and "borrowed genes" depends on how far back you look.

Barriers: why cells do not become a free-for-all

If DNA moves so easily, why do bacterial species exist at all? Several barriers limit transfer, and they are worth knowing, because they explain why HGT is common but not unlimited.

• Sequence similarity. Incorporating transformed DNA into a chromosome usually requires that the incoming DNA is similar to a segment already there. DNA from a distant relative may be taken up but not integrated.
• Restriction-modification systems. Many bacteria carry enzymes that cut incoming DNA at specific sequences while protecting their own DNA with chemical tags. These act like a security check at the door.
• CRISPR systems. Some bacteria and archaea store snippets of past phage DNA and use them to recognize and destroy matching invaders. The technology that made CRISPR famous in the laboratory began as a defense against exactly the kinds of transfer discussed here. See CRISPR gene editing in 2026 for the engineering side of that story.
• Cost. A new gene is not free. It has to be copied each time the cell divides, and if it is not useful, selection will remove it. The 18% figure from the E. coli study reflects what survived this filter.
• Compatibility. A gene must be turned on in the right way by the host's machinery to do anything. Genes from distant relatives often arrive silent or unregulated.

The upshot is that HGT is a constant supply of candidate genes that selection filters. Most are neutral or harmful and are lost. A small fraction help in some environment, and over time that fraction accumulates.

Why microbes benefit, and why we notice

What would a gene need to do to be worth keeping? Cases that show up repeatedly in the literature include genes for resistance to antibiotics and heavy metals, genes that break down unusual food sources, genes for toxins and adhesion factors that make a bacterium a better pathogen, and genes that help a cell survive stress. In each case, the gene comes with a ready-made function that is hard to evolve from scratch.

Researchers who study environmental samples report that the pace of transfer responds to conditions. In one summary, textile wastewater contained 3- to 13-fold higher abundance of mobile genetic elements than other wastewater samples, and some phenolic compounds raised HGT rates 2- to 9-fold in experiments. These are results from specific studies and do not describe all environments, but they support the general idea that pollutants and stress can increase the amount of gene movement.

The human microbiome, the community of microbes in and on the body, is also a venue for HGT. Soucy and colleagues's review lists it among the settings where transfer matters. For readers interested in the community itself, microbiome research explains what lives there and what it does.

What people get wrong about HGT

"It means evolution is not gradual." HGT does not contradict natural selection. It supplies new genetic variation by a different route, and selection still decides what stays. The framework in evolution and natural selection applies unchanged. What changes is the source of the variation: copying errors, recombination, and imports.

"Bacteria are constantly swapping genes with humans." The cases in animals are limited and, as the tardigrade episode shows, often contested. Gene movement from bacteria into human germ-line cells, the kind that would be passed to children, has not been convincingly demonstrated. Reports of bacterial sequences in human DNA should be treated cautiously, especially when they come from cancer samples or contaminated assemblies.

"HGT makes a species concept meaningless." It complicates it. Many bacteria still fall into recognizable groups with a stable core genome, and the groups persist for long periods. The concept of species is less sharp for microbes than for birds, but that was true before HGT was discovered.

"It only matters for antibiotics." Antibiotic resistance is the most visible consequence, because it has a direct effect on medicine. In evolutionary terms, though, HGT also moves genes for ordinary metabolism between lineages, including the genes that let a microbe use a new food source or break down a pollutant.

Frequently asked questions

Q: What is horizontal gene transfer in simple terms? A: It is the movement of genes between organisms that are not parent and offspring. Instead of passing DNA down to children, a cell copies a piece of its DNA into another cell, which can belong to a different species. Bacteria do this regularly through transformation, conjugation, and transduction.

Q: How is horizontal gene transfer different from vertical gene transfer? A: Vertical transfer is the ordinary inheritance of genes from parent to offspring when a cell divides or an organism reproduces. Horizontal transfer moves genes sideways between unrelated individuals, so a new gene can appear in a lineage without being inherited from its ancestors.

Q: Does horizontal gene transfer happen in humans? A: Not in the way it does in bacteria. Reports of foreign genes in the human genome, such as one proposal of about 100 genes out of roughly 20,000, have been disputed. Humans do carry large amounts of DNA from ancient viral insertions, but that is a separate process from the swaps described here, and the evidence for bacterial genes in the human germ line is weak.

Q: How does horizontal gene transfer cause antibiotic resistance to spread? A: Resistance genes often sit on plasmids or other mobile elements. When a resistant bacterium passes a plasmid to another cell by conjugation, or a phage carries the gene by transduction, the recipient becomes resistant without having been exposed to the drug. A single plasmid can carry resistance to several antibiotics at once, as the R factors discovered in Japan in 1959 showed.

Q: Who discovered bacterial conjugation? A: Joshua Lederberg and Edward Tatum reported it in a letter to Nature on October 19, 1946, after mixing two nutritionally deficient strains of E. coli K-12 and finding colonies that could grow on a minimal medium. Lederberg shared the 1958 Nobel Prize in Physiology or Medicine for his work on genetic recombination in bacteria.

Q: Has horizontal gene transfer changed the tree of life? A: Yes, in how it is drawn. For microbes, trees built from different genes often disagree, and many biologists describe the relationships as a web rather than a single branching tree. For animals and plants, vertical inheritance still dominates, and the tree remains a good model.

Q: Can viruses move genes between bacteria? A: Yes. The process is called transduction. A bacteriophage occasionally packages host DNA into a new virus particle and delivers it to the next cell it infects. In Vibrio cholerae, the genes for cholera toxin are carried by a phage, so infection can turn a harmless strain into a toxin producer.

The takeaway

The standard picture of heredity, with genes flowing down a line from parent to child, is accurate for a large part of life and incomplete for the rest. In bacteria and archaea, genes also flow sideways, by three main routes that were uncovered between 1928 and 1952, and by a handful of rarer ones that continue to turn up. The result is a world where a harmless cell can become a pathogen by receiving one package, where a drug-resistance gene can cross species lines in an afternoon, and where the history of a genome reads more like a patchwork than a pedigree.

The most useful habit this subject teaches is to separate two questions that are easy to blur. One is "how are these organisms related?" The other is "where did this particular gene come from?" For most animals the answers match. For microbes they often do not, and for each claim of a surprising transfer the right response is to ask for the evidence. The tardigrade correction shows that the same method that finds real transfers can also produce false ones. Careful reading is part of the science.

Sources

• Ochman, H., Lawrence, J. G., and Groisman, E. A. (2000). Lateral gene transfer and the nature of bacterial innovation. Nature 405:299-304. PubMed record
• Lawrence, J. G., and Ochman, H. (1998). Molecular archaeology of the Escherichia coli genome. PNAS 95:9413-9417. PubMed record
• Lederberg, J., and Tatum, E. L. (1946). Gene recombination in Escherichia coli. Nature 158:558. Catalog record
• Waldor, M. K., and Mekalanos, J. J. (1996). Lysogenic conversion by a filamentous phage encoding cholera toxin. Science 272:1910-1914. DOI
• Gladyshev, E. A., Meselson, M., and Arkhipova, I. R. (2008). Massive horizontal gene transfer in bdelloid rotifers. Science 320:1210-1213. DOI
• Kyndt, T., et al. (2015). The genome of cultivated sweet potato contains Agrobacterium T-DNAs with expressed genes. PNAS 112:5844-5849. DOI
• Koutsovoulos, G., et al. (2016). No evidence for extensive horizontal gene transfer in the genome of the tardigrade Hypsibius dujardini. PNAS 113:5053-5058.
• Watanabe, T. (1963). Infective heredity of multiple drug resistance in bacteria. Bacteriological Reviews 27:87-115.
• Soucy, S. M., Huang, J., and Gogarten, J. P. (2015). Horizontal gene transfer: building the web of life. Nature Reviews Genetics 16:472-482. Nature
• Centers for Disease Control and Prevention. About antimicrobial resistance
• Overview articles used for dates and summary figures: Horizontal gene transfer, Bacterial conjugation, Transduction, Griffith's experiment