A new computational pipeline developed at the University of Maryland School of Medicine may help improve malaria vaccine development by addressing two major challenges: the genetic diversity of Plasmodium falciparum and variability in human immune genes across African populations. The model was presented at the 2025 Annual Conference on Vaccinology Research.
The approach uses mixed-integer linear programming to identify T cell epitopes that are conserved, immunogenic, and broadly recognizable across a range of human leukocyte antigen (HLA) alleles. The tool evaluates parasite sequences from 18 African countries and 748 unique HLA alleles from 24 African nations. It then selects combinations of epitopes that maximize population-level immune coverage while reducing the likelihood of immune escape.
From a set of 42 candidate proteins, the model identified more than 65,000 class I and nearly 2,000 class II epitopes. The most promising liver-stage antigens included HSP70-2, ROM1, and LISP1. Blood-stage antigens included RON2, PfRh5, and EBA-175. Among the most optimized results, five MHC class I epitopes independently covered 100 percent of HLA-A, -B, and -C alleles. MHC class II coverage ranged from 98.5 to 100 percent when epitopes from multiple parasite stages were combined.
Alexander Laurenson, a second-year MD/PhD student and lead investigator, said the pipeline aims to help researchers narrow their focus before entering more costly stages of vaccine testing.
“What I think we’re experiencing right now in a lot of the malaria vaccine development pipeline is a bottleneck in determining what exactly to target,” Laurenson said. “Plasmodium falciparum is a eukaryote, so it has an enormous genome. That means there are a lot of components that allow for redundancy in how the pathogen invades, causes disease, and survives in the host.”
Laurenson noted that by using this tool early in the vaccine development pipeline, researchers can reduce time spent on expensive laboratory assays.“We’re hoping this tool is integrated effectively so we can at least generate a shortlist of potential antigens or epitopes to target before moving on to that expensive and tedious step,” he said.
The tool also focuses on enhancing HLA coverage, a factor tied to the ability to activate T cells, which support longer-lasting and more durable immune responses.“When we talk about HLA coverage, we’re covering a much more complicated concept related to the activation of T cells,” Laurenson said. “This is something that’s often less emphasized in vaccinology, where the focus is usually on generating a robust antibody response—which is, rightfully, considered highly protective.”
“But in this case, we’re focusing on what I sometimes call the ‘middle child’ of the immune response—T cells—that help do a few important things. CD8 T cells go after infected host cells to kill them and squash the infection when it’s hiding inside our cells. CD4 T cells, when activated, fully activate B cells, which is key for those B cells to become memory B cells and produce long-lasting antibodies.”
By improving T cell engagement, the model may contribute to longer-lasting immunity and potentially reduce the frequency of booster doses needed in the field.
Laurenson said the tool was designed to be scalable and adaptable to other vaccine targets. While its initial application was malaria, he emphasized its potential for other diseases with similar challenges in antigenic diversity and immune variability.
“We’re not just in the business of designing malaria vaccines—we’re also interested in designing tools that could be used for other understudied, neglected, or elusive pathogens,” he said. “We’re looking at applying this work to pathogens that have been difficult to target with vaccines due to high antigenic diversity, or where there’s variability in immune response between individuals.”
He noted that this includes diseases like influenza, COVID-19, and hepatitis B, where breakthrough infections remain common and vaccine coverage varies.“We really want to figure out how this overall strategy could be used not just for malaria, but for other emerging or elusive pathogens,” Laurenson said. Next steps include experimental validation of the model’s epitope predictions and integration into vaccine platforms that support durable immune responses.
