One metagenomic test for detection of all viruses at once

Ellen Carbo received her PhD from Leiden University on May 17 for her thesis entitled ‘Metagenomic sequencing in clinical virology: advances in pathogen detection and future prospects’. Prof. ACM Kroes acted as supervisor. Dr. JJC de Vries and Dr. IA Sidorov were co-supervisors. Carbo works as a medical molecular microbiologist in training at the Amsterdam UMC and is also coordinator of the special interest group bioinformatics for the Dutch Association for Medical Microbiology (NVMM).

Photo: Froukje Vernooij

What was the purpose of your PhD research?

To detect viruses, modern diagnostics searches specifically for one or more specific viruses. This is often done by using a targeted polymerase chain reaction (PCR) test. However, to do this, one must already have a suspicion of which virus is expected in a patient. However, the suspicion is often incorrect or one has no idea which virus is expected. The aim of this PhD research was therefore to set up one metagenomic test to detect all viruses at once directly from a patient sample. This test maps the entire metagenome by processing all genetic material directly from a patient sample and analyzing it for the presence of viral microorganisms.

What do you want the clinical doctor to know about your research?

Metagenomic sequencing offers the option of identifying the causative agent in infections where no pathogen is found with regular diagnostics. In my PhD research I only looked at viruses, but other microorganisms can also be mapped with the metagenomic test, without the need for a culture. The test is not cheap and still takes a relatively long time, but if it is further improved, it can be used much more often in patients where there is a suspicion of infection but where diagnostics for a pathogen remain negative. As a result, a pathogen will be found in more patients, allowing for more adequate treatment.

What was the most frustrating part of your research?

Patient material submitted often consists of less than 0.1% viruses. This makes searching for any virus present very difficult; it’s basically like looking for a needle in a haystack. In many patient materials, the majority of the genetic material consists of human DNA and RNA, between which you have to find your viral material and with reliable sensitivity.

What moment/insight brought a breakthrough?

What has greatly improved the sensitivity of our test is – after processing samples for next-generation sequencing (NGS) – to select the viral material from a sample using probes before sequencing. This led to us having 100 to 10,000 times greater coverage of our viruses.

These probes have thousands of viral targets, including for viruses that can infect vertebrates. The probes also allow for some mismatches, which means that some mutations in a virus genome would not immediately lead to a negative test. These properties of the probes also meant that we immediately had a working sequencing protocol at the beginning of the COVID-19 pandemic. Our probe pool contained probes for bat corona sequences, but this also allowed us to perfectly map the SARS-CoV-2 genome.

It is desirable to have the metagenomic sequence test in different laboratories, because it is also an important addition for surveillance and monitoring in the event of new outbreaks.

What is the follow-up question that arises from your research?

It would be interesting if we could make the test even more accurate without using probes and thus map all the genetic material, both of the host and of the micro-organism. The follow-up question would then be whether we could also use the metagenomic test to determine the causality of a micro-organism by means of transcriptome sequence analysis, instead of the current diagnostics in which we actually only determine whether a micro-organism is present.

We could also use the metagenomic sequence data to look at genetic properties of the microorganism, such as resistance to medication, or properties of the host itself, such as pharmacogenetic profiles. We could also use the metagenomic sequence data for oncology and any genetic properties that influence inflammation. A metagenomic test may be something from microbiology, but in the future it could be used so much more broadly than just looking for micro-organisms. The metagenomic test therefore has important future prospects.

What do you take away from your PhD research?

There is still quite a lot unknown about micro-organisms, for example many genomes of micro-organisms have not yet been fully developed. With the complete field of work, we must ensure that micro-organisms and their functioning are further mapped out, which will ultimately lead to better patient care with regard to infections. What I find very important is that we make full use of the amount of sequence data we generate, but also that we process this sequence data reliably. This is therefore a reason for co-coordinating a knowledge network for bioinformatics within medical microbiology.