Peptone, a London-based startup that is using artificial intelligence and advanced atomic imaging techniques to pursue a tricky class of potential targets for new medicines, has raised an additional $40 million in venture capital investment to fund its expansion.

F-Prime Capital, a venture capital firm that specializes in health care and technology investments, and Bessemer Venture Partners lead the new investment round, with participation from Walden Catalyst Ventures. Peptone’s existing investors, London-based Hoxton Ventures, and dRX Capital, the venture arm of pharmaceutical company Novartis, also participated in the new financing. This Series A funding—the first large institutional financing round for a startup—brings the total funding Peptone has raised since its founding to about $42 million.

Recently, researchers have made tremendous advances in using artificial intelligence to predict the complex three-dimensional shapes of proteins. Proteins are long strings of amino acids that form the building blocks of most biological processes. When in isolation, they spontaneously fold, or collapse, into highly-complex three-dimensional shapes. But until recently, scientists had few ways to predict what these shapes would be. The breakthrough in using A.I. to forecast these structures with a high degree of accuracy based on knowing the protein’s genetic code is seen as potentially speeding the discovery of new drugs because many drugs are small molecules that bind to a specifically shaped “pocket” in the structure of a protein.

But as many as 50% of human proteins either don’t fold into regular shapes, or contain regions that don’t have a clear structure—at least not when the protein is found in isolation. Scientists call such proteins “intrinsically disordered.” Yet just because these proteins don’t seem to have a consistent shape doesn’t mean they don’t do anything. Scientists believe these disordered proteins and disordered protein regions play a critical role in biological processes, including key aspects of cell function and cell reproduction, and in many diseases, including Alzheimer’s, cancer, and diabetes. Yet researchers’ inability to predict the protein’s structure has made it difficult to find drugs that will bind with these proteins.

Peptone, which was founded in 2018, has developed a potentially game-changing method to predict the way disordered proteins are likely to behave in the presence of other proteins and molecules. The startup, which has offices in London and in Bellizona, Switzerland, uses advanced imaging techniques that are able to analyze the dynamics of disordered proteins in a fluid state. Peptone then uses this data to create computer simulations of how a disordered protein will interact with other molecules at an atomic level. These computer simulations require the use of a specialized high-performance computer. The company uses artificial intelligence techniques to search for molecules that can bind with these disordered proteins. And then it validates these molecular candidates using traditional lab experiments.

Kamil Tamiola, Peptone’s founder and chief executive officer, told Fortune that the new funding will allow Peptone to complete the construction of its laboratory facilities in Bellizona, a biotechnology hub in southern Switzerland. Peptone needs to purchase and install expensive, sophisticated machines to perform the two kinds of imaging—nuclear magnetic resonance (NMR) spectroscopy and hydrogen-exchange mass spectrometry—it uses to analyze the shapes disordered proteins assume as they move. The more commonly used ways methods of protein imagery, X-ray crystallography and cryo-electron microscopy; both involve turning proteins into crystals that lock their structures in place. Disordered proteins are, by their very nature, not amenable to these techniques because their structures are fluid, and vary depending on the presence of other proteins and molecules.

Tamiola said the new funding will also allow Peptone to investigate possible drug candidates for five different disordered protein targets that it has identified as promising and taking them through to pre-clinical validation. He said he expects this process to take about 14 months. The company also has two pilot projects with large pharmaceutical companies, which he declined to name, citing nondisclosure agreements.

“Modulating disorder is our big thing,” Tamiola said, noting that disordered proteins often experience moments when, in the presence of the right other protein or molecule, they adapt a more stable structure. In that condition they become potentially druggable. “We want to say you can design an artificial peptide that maybe intuitively makes no sense, but if you put it in the cell it will bind,” he says. A peptide is a small chain of amino acids, whereas a protein is a much longer chain composed of many peptide groups.

Francis Ho, a partner at Walden Catalyst Ventures, said in a statement that his venture firm believes that by uncovering ways to make disordered proteins druggable, Peptone could fill a crucial gap that has been left unaddressed by other A.I.-enabled drug discovery companies.

Peptone’s technology has “vast potential” to treat major diseases, Andrew Hedin, a partner at Bessemer Venture Partners, said in a statement.