The paper, authored by researchers at the French biotechnology company NEVEZYNE SAS and the Kenya Medical Research Institute (KEMRI), lays out a dual-track strategy. On one track, machine learning compresses the discovery of new malaria antigens and antibodies from years to months. On the other, the silkworm Bombyx mori, infected with engineered baculoviruses, serves as a living bioreactor that can churn out those molecules at a fraction of the cost of conventional mammalian cell culture. The authors argue this combination directly supports the African Union’s goal of manufacturing 60% of the continent’s vaccines locally by 2040.

The scientific case for AI in malaria research is already substantial. The WHO-approved vaccines RTS,S/AS01 and R21/Matrix-M target the parasite’s circumsporozoite protein (CSP), but their protection wanes and falls well short of sterilizing immunity. Machine learning is now being used to find better targets: a 2024 study applied positive-unlabeled learning to 272 variables across the Plasmodium falciparum proteome, prioritizing conserved antigens with low genetic diversity that could elicit strain-transcending immunity. Meanwhile, deep learning tools like AlphaFold2 have resolved the structures of key targets such as the transmission-blocking protein Pfs48/45, and a 2025 study in PNAS used machine learning to design multi-epitope scaffolds presenting multiple copies of the PfCSP epitope in optimally oriented configurations—antigens that outperformed the native protein in preclinical models.

AI is also reshaping the antibody side of the pipeline. The monoclonal antibody CIS43, which binds a vulnerable junctional region of PfCSP, showed up to 88.2% sterile protection over a six-month malaria season in a phase 2 trial in Mali when given as an engineered half-life-extended variant called CIS43LS. A second antibody, L9LS, targeting a distinct conserved epitope, achieved 77% efficacy against clinical malaria in a phase 2 trial among Kenyan children. Deep learning models can now analyze the structural interactions between these antibodies and their epitopes, predicting substitutions that boost binding affinity, broaden cross-strain reactivity, or improve stability—accelerating the design of next-generation antibody cocktails that attack the parasite at multiple life stages simultaneously.

But designing a molecule on a computer is only half the battle. The authors emphasize that current biomanufacturing is structurally ill-suited to the malaria burden. Mammalian cell culture facilities, the gold standard for producing recombinant proteins and monoclonal antibodies, require more than €50 million in capital expenditure and take nine to twelve months to move from gene sequence to a GMP-grade batch. mRNA platforms, while fast to develop, demand ultra-cold storage and complex lipid nanoparticle manufacturing that few countries can support. The result is what the authors call a geography of dependence: sub-Saharan Africa accounts for 94% of global malaria cases, yet not a single malaria vaccine or anti-malarial monoclonal antibody is manufactured on the continent.

The silkworm platform, known as the Bombyx mori baculovirus expression vector system (BEVS), offers a radically different economics. Instead of stainless-steel bioreactors, the living insect itself becomes the production unit. A silkworm-based facility is estimated to cost between €2 million and €5 million—roughly a tenfold reduction in capital expenditure—and can theoretically deliver a GMP-ready product within 100 days of sequence availability, aligning with the CEPI 100 Days Mission. A single silkworm pupa yields the equivalent recombinant protein of 100 to 1,000 milliliters of conventional insect cell culture, and because each insect is an independent unit, the system scales out linearly rather than requiring ever-larger tanks.

Crucially, the platform is not limited to simple antigens. Silkworm larvae possess a eukaryotic secretory pathway capable of proper disulfide bond formation and glycosylation, meaning they can assemble fully functional human IgG antibodies. In a landmark 2021 study, researchers produced the SARS-CoV-2 antibody CR3022 in silkworm larvae as scFv, Fab, and full-length IgG formats, with binding affinities nearly identical to mammalian-produced versions. Yields can reach up to 36 micrograms of recombinant IgG per larva, and the authors project that the cost per dose could drop from roughly €3–5 for mammalian systems to under €1 for silkworm-derived products.

The regulatory pathway is also more mature than the novelty of the idea might suggest. Several human vaccines produced in insect cell BEVS—including GSK’s Cervarix, Sanofi’s FluBlok, and Novavax’s Nuvaxovid—have already been approved by major agencies such as the EMA and FDA, establishing a safety precedent for baculovirus vectors and insect-derived glycosylation. For whole-insect systems specifically, the Spanish company Algenex has secured EMA approval for veterinary vaccines using Trichoplusia ni pupae, and the Japanese firm KAICO is advancing a silkworm-derived human norovirus vaccine into phase 1 clinical trials in close consultation with Japan’s PMDA. The authors argue African regulators can align with these precedents through WHO prequalification pathways.

What makes the proposal genuinely novel is the supply chain it envisions. Rather than building pharmaceutical infrastructure from scratch, the authors propose a farm-to-GMP model in which Africa’s existing sericulture sector—already established in Kenya, Uganda, Rwanda, and Ethiopia with support from institutions like ICIPE and KALRO—supplies healthy silkworm larvae to centralized GMP facilities. Farmers handle only the agricultural side, rearing larvae under conditions identical to those used for textile silk production. The larvae are then transported to specialized laboratories where baculovirus infection and protein purification occur under controlled pharmaceutical conditions. This creates a high-value cash crop for rural communities while building pharmaceutical manufacturing capacity, addressing economic development and health sovereignty in a single intervention.

The timing is significant. The African Union launched the Partnerships for African Vaccine Manufacturing (PAVM) in 2021, and Gavi’s African Vaccine Manufacturing Accelerator committed over $1 billion over ten years in milestone-based financing. Yet a 2025 Lancet Global Health analysis found only eight African countries have achieved WHO regulatory maturity level 3, and most existing manufacturers are limited to fill-and-finish operations. Kenya’s experience illustrates both the promise and the peril: President William Ruto has committed to producing 50% of the country’s essential health products domestically, and the Kenya BioVax Institute is developing the nation’s first dedicated vaccine facility. But Moderna paused a $200 million mRNA plant in Kenya in 2024, citing insufficient long-term demand signals—a cautionary tale about high-capital bets. The authors argue the silkworm platform’s low capital threshold makes it viable even without guaranteed procurement contracts, and Kenya’s established sericulture sector provides an immediately available biological supply chain. The economic stakes are enormous: the African Leaders Malaria Alliance warns a malaria resurgence could cost sub-Saharan Africa up to $83 billion in lost GDP by 2030, while sustained elimination efforts could yield a $230 billion boost. By combining AI’s ability to design highly efficacious, strain-transcending antigens with a manufacturing platform whose costs are an order of magnitude lower, the authors contend the return on investment for malaria elimination fundamentally changes—making next-generation vaccines and antibodies economically feasible for the countries that need them most.

Subject of Research: AI-driven design of malaria vaccines and monoclonal antibodies combined with silkworm-based biomanufacturing to enable local production in Africa

Article Title: AI and silkworm biomanufacturing: a new paradigm for malaria therapeutics access

Article References: Pilorget, A., Njeru, S. N., Matoke-Muhia, D., & Loizel, C. (2026). AI and silkworm biomanufacturing: a new paradigm for malaria therapeutics access. Frontiers in Malaria, 4, Article 1909034. https://doi.org/10.3389/fmala.2026.1909034

Image Credits: AI Generated

DOI: 10.3389/fmala.2026.1909034

Keywords: malaria, artificial intelligence, silkworm, biomanufacturing, vaccines, monoclonal antibodies, Plasmodium falciparum, Africa, vaccine manufacturing, PAVM, sericulture, global health