Infection diagnostics under the microscope with metagenomics

“A boy in Salt Lake City was hospitalized with symptoms of pneumonia. He was given antibiotics and the sample collected was sent to the lab for a PCR test.” The test turned out to be positive for a season of coronavirus, but that did not explain the serious symptoms. “The boy recovered and was sent home. Three days later he returned with worsening symptoms. Pleural fluid was then collected and that sample ended up in our lab for analysis metagenomic analysis.”

In the IDbyDNA lab, Rossen was able to map all microorganisms in the sample using clinical metagenomics. There they found Prevotella pleuritidis, a fairly rare bacterium. “We had not thought about that in advance,” says Rossen. “It has not been known for very long that this bacterium can cause serious respiratory symptoms. The antibiotics the boy had already received were not suitable for this anaerobic bacteria and after the correct antibiotics were prescribed he recovered.”

Non judgmental

When diagnosing an infectious disease, the clinical picture is used to screen for microorganisms that match that picture. It becomes more complicated when you are dealing with an unknown disease or a complicated clinical picture. In that case you don't know what you are looking for. Metagenomics allows us to analyze the entire genetic material in a sample. Rossen: “With metagenomics you look at a clinical sample in an unbiased way. By analyzing a sample in this way, in some cases you find something that you had not thought of beforehand and probably had not looked for.”

Rossen notes that culture is in principle also an unbiased way of looking at a sample, but that this has a number of major disadvantages compared to metagenomics. It takes a long time to culture a bacteria or virus and a diagnosis can take days or sometimes even weeks. Some bacterial species are also not easy to grow, such as the anaerobic bacteria in the example above.

Metatranscriptomics

Metatranscriptomics goes one step further and allows researchers to not only determine the presence of a bacterium or virus, but also say something about its behavior. “We use metatranscriptomics to look at what exactly happens. For example, you may determine that a certain bacteria is present, but that does not mean that it is the cause of the infection. We also look at the host response.”

Sequencing the entire genetic material produces a huge amount of data. You therefore need smart software to solve this puzzle. Rossen: “We used an AI model to screen information about certain bacteria or viruses in a curated database. The model compares information from the database with the patient's results and symptoms.”

Standardization

The possibilities of genomics technologies are not limited to diagnostics. Rossen, for example, sees opportunities in the preventive screening of wastewater in hospitals to make a prediction about infectious disease outbreaks. It is also used to identify new pathogens, such as SARS-CoV-2. But according to Rossen, further standardization is needed to make the method widely applicable. Location is an important consideration: “The question is whether you use this for all patients, or only for seriously ill patients, or only when all other tests are negative. There is no clear answer to this yet. Above all, we need more data to see what the value of this technology is. I hope that more research will be done into this, and that it will also be taken into account what benefits the patient.”