Researchers in the UNMC Department of Pathology and Microbiology published a paper this week that points to a faster, less labor-intensive screening protocol to help screen for antimalarials.
Caroline Ng, PhD, an associate professor of pathology and microbiology, and recent PhD graduate Melissa Rosenthal, PhD, published their research on the value of high content imaging and machine learning in the national journal Cell Reports Methods. The paper is titled, “High-content imaging as a tool to quantify and characterize malaria parasites.”
The research specifically addresses the parasite Plasmodium falciparum, which is the parasite responsible for causing severe cases of malaria. These parasites have a relentless capacity to acquire drug resistance, which presents a major hurdle for eliminating malaria.
Current protocols to identify drug effects on malaria parasites call for using flow cytometry and light microscopy. Flow cytometry only detects a signal from labeled parasites, so researchers cannot see the morphology of parasites treated with drugs. Light microscopy allows morphological features to be visualized but is laborious and time-consuming. The UNMC researchers’ new method combines the positive attributes of both these methods and leaves behind the negative aspects.
The researchers used high-content imaging on the Operetta CLS imager, which they noted has the capability to rapidly collect a multitude of properties used to quantify cells and cell phenotypes.
Then, the researchers leveraged the power of machine learning – through the software PhenoLOGIC – to analyze the phenotypic data generated from the high-content imaging.
After a human initially classifies objects to study as an output, the software learns from that training set and identifies which combination of properties are important for classifying objects, according to the research. The method should be accessible to biologists with limited technological backgrounds, the paper notes, because it does not require expertise in coding or machine learning. In addition, this method reduces person-to-person bias in parasite identification and reduces required training of new personnel.
The approach identified red blood cells infected with malaria parasites. The data analysis was able to distinguish parasite development stages and automatically count the number of nuclei within a parasite, according to the paper.
The research concluded that “the counts from the machine learning approach were nearly identical to previously used methods of flow cytometry and light microscopy.” That indicates this new method is as accurate as previously used methods, the authors say.
This approach can be used to screen for antimalarials and identify at which parasite developmental stage the drug acts on, which can inform how health care professionals design antimalarial combination therapies, because they want drugs that act on different parasite stages so all are effectively inhibited.
In the future, the research says that continual development of antimalarials will be essential to overcome a chronic state of drug resistance. The duo of high-content imaging with machine learning is a valuable tool in this endeavor, the paper says, by rapidly identifying potent compounds and providing clues toward the mode of action of these compounds.
Dr. Ng said she hopes this method will be another tool in the arsenal to identify antimalarial compounds and their effects on malaria parasites.
Way to go Caroline! Keep up the good work!
Congratulations to Dr. Ng and Rosenthal on this work!!