Buggin’ out with computers

Alder Zomer
Aldert Zomer

When I did my PhD, it took two years to sequence and assemble the genome of a single bacterial strain. I have to learn bioinformatics, I thought. Today’s DNA sequencing output and computational capacity have advanced so much in an incredibly short time. Instead of sequencing individual bacteria, we can now study many species within their natural environment using metagenomics.

My group and I use bioinformatics to study antibiotic resistance in bacteria and the transfer of bacteria between humans (called zoonosis) and animals using molecular epidemiology and metagenomics. I work closely with several wet labs which means that all of our data is newly generated in the lab, and that we can go back to our lab partners to test our computational predictions.

Medicinal poop

One of our metagenomics projects, run by PhD student Mathijs Theelen, is using fecal transplantation to treat colitis and chronic diarrhea in horses. Bacteria from a stool sample of a healthy donor horse are administered to the recipient horse by nasogastric tube (a tube that goes in through the nose into the stomach), where they restore the balance in the intestinal microbiota. This technique has been successfully used in horses for years and in human medicine has recently gained interest, especially in the treatment of Clostridium difficile infection, a bacteria that causes severe diarrhea and inflammation of the intestine. Yet no one knows exactly how this treatment works. Do the good bacteria colonize the intestines, restoring the balance, or do they actively get rid of infectious agents?

Vaccine development

Together with Janssen Pharmaceuticals, we’re developing vaccines against invasive E. coli bacteria in humans. E. coli is one of the most common pathogens in humans and quickly gains multidrug resistance. We’re studying genome sequences, to figure out how different bacterial strains from patients are related and investigate the variation in the parts of the bacterium that are used as vaccine targets. With this information, we can adjust the vaccine to make it more effective and potentially find new targets for improved vaccine composition.

We work together with computational scientists and wet lab scientists, with farm owners and pet owners, with clinicians, both human and veterinary

Transfer of resistance

Salmonella is also a common pathogen, often associated with chickens and chicken products and is easily transferred to humans, generally causing food poisoning. Although many cases of salmonella infection resolve themselves, a number of cases require antibiotics. Similar to E. coli, resistance to antibiotics is common. Salmonella carry their resistance on plasmids, a small piece of circular DNA that can easily be transferred from one resistant bacteria to other bacteria, including E. coli. A project led by PhD candidate Ricardo Castellanos Tang focuses on the relatedness of resistant strains and their plasmids to map out the development of resistance, and we have developed software to investigate transfer of resistance via plasmids. 

Identifying the cheesy culprit

Another bacterium, Campylobacter fetus (C. fetus) lives in the gastrointestinal tract of cows and sheep. Although very rare in humans, C. fetus can cause a dangerous invasive infection, potentially requiring life-long antibiotics. In 2015, there were 3 cases of C. fetus infection in a very short time frame and within the same province. Soon after, 2 additional cases were reported. This was considered an outbreak and mobilized our team very quickly.

We were able to identify two sheep farms as the possible sources of infection: all patients had eaten unpasteurized sheep’s milk cheese, and one farm housed animals that tested positive for C. fetus. Traditional methods of identification showed that the bacterial strains from the positive farm and unrelated cases were the same and could not be distinguished from each other. So, we investigated further using genome sequencing. And indeed, our results showed a distinct bacterial strain with a genome identical to the patient isolates, carried by the sheep from the positive farm allowing us to pinpoint this farm as the source of the outbreak. This a good example of how genomics can distinguish subtle differences that can quickly stop the spread of infection and prevent additional (economic) consequences.

What I find most inspiring is that bioinformatics is collaborative. We work together with computational scientists and wet lab scientists, with farm owners and pet owners, with clinicians, both human and veterinary. This ‘One Health’ approach makes Utrecht a great place to combine traditional laboratory work with novel bioinformatics.