Every year, hepatitis C virus affects more than 200 million people worldwide. Infection is the main cause of chronic liver disease, including chronic hepatitis, cirrhosis, and liver cancer.
Unlike hepatitis A and B, there are no approved vaccines for hepatitis C.
Hepatitis C antiviral drugs have revolutionized treatment with high success rates, but there is another class of highly effective drugs, especially in drug resistance and disease prevention.
These drugs are short chains of amino acids and are based on highly specific peptide molecules targeting viral components, improving solubility and lower toxicity. It can also work on any genotype of the virus, making it easier to synthesize than traditional small molecule drugs.
However, the problem lies in searching for peptides that function against certain viruses. This is because there are many peptides, but very few function against certain viruses. This is a similar process to finding a needle in a haystack.
Recent research has shown that using machine learning is an efficient approach to searching for these molecules.
Machine learning is used to predict common antibiotics and antiviral peptides, but only a handful of studies focused on research targeting specific viruses such as hepatitis C.
Find the peptide
Proteins and peptides are given great attention from scientists and researchers around the world regarding their potential for use in diagnostic, therapeutic, and drug delivery systems.
For example, insulin is a peptide containing 51 amino acids, and its discovery and development is considered to be one of the most important advances in drug discovery. Recently, semaglutide, a Waitros drug, which is a peptide containing 31 amino acids, has been hailed as a game-changer for patients with type 2 diabetes, obesity and heart disease.
Both proteins and peptides are chemically the same. Peptide chains are made up of amino acid residues, the same chemical molecules that make up proteins and allow cells to perform a variety of functions.
Given the natural presence of 22 amino acids, it is possible to create distinct short peptide chains in a variety of ways. For example, synthesis of a peptide chain consisting of 10 amino acid residues could result in a large amount of (22¹⁰) of different chains.
However, a very small number of these peptides are therapeutically effective and even less likely to act on certain viruses.
Moreover, the number of FDA-approved bioactive peptides (those with therapeutic properties) is very limited. Therefore, identifying the ones that exhibit antibacterial and antiviral properties among the vast possibilities is Hercules' challenge.
Here, machine learning helped researchers.
Traditional methods for identifying effective antihepatitis peptides, or drug molecules in question, are time-consuming and resource-intensive.
Some estimates suggest that traditional methods require synthesis of up to 5,000 molecules over four to six years to find promising candidates for drug development.
However, machine learning models can significantly reduce screening times. They can screen billions of molecules, allowing researchers to process one billion molecules every day. This makes it quick and cost-effective, especially in the early stages of drug development.
One such model that researchers have recently developed is a web-based prediction tool called Pred-AHCP (predicting anti-hepatitis peptides) to assess whether peptide molecules can effectively inhibit hepatitis C virus.
This model does this by analyzing its amino acid composition and physicochemical properties.
This method employs a two-stage computational filtering process that relies on statistical algorithms. This approach is particularly useful because it predicts whether a peptide may be antihepatitis, as well as explains the reason for its effectiveness by highlighting the most important molecular properties of the molecule.
Therefore, rather than simply predicting candidates, researchers can understand the underlying mechanisms that may make them work.
Virus-specific prediction tools
Beyond treatment of hepatitis C, this approach can be adapted to develop similar predictive tools for searching for peptides against other viruses, thus creating a family of virus-specific predictive tools.
This is especially valuable for viruses that currently lack effective treatments such as HIV, herpes simplex virus, and Zika.
Discoveries from the development of Pred-AHCP can facilitate more efficient discovery of lead peptide candidates than common antiviral peptide prediction methods.
This provides insights that molecular features can make certain peptides effective against hepatitis C, due to the explanatory nature of this machine learning model, which provides synthetic biochemists, organic chemists and bioengineers.
Such an understanding can lead to the rational design of new therapeutic agents by highlighting essential properties such as hydrophobic distribution. This is an important physicochemical property that is often used in the presence of drug design, toxicology, environmental monitoring – the presence of specific amino acid pairs, and other structural elements.
For example, modifications derived by the model can be tried to maximize the effect of key features, improving potentially effectiveness, sustained virological responses, and even pharmacological properties such as stability, permeability, easy drug properties to the extent that they diffuse across the biofilm, and even bioeigability (in fact, some of the absorbents are absorbed).
Additionally, this model is available as a web server, allowing greater accessibility among researchers without specialized computational expertise.
By sharing therapeutic approaches, such democratization of advanced computational tools could enhance collaboration and innovation in antiviral research for a variety of infectious diseases.
It has been originally published Creative Commons 360INFO™.
Akash Saraswat He is a senior researcher at BML Munjal University in Gurugram, Haryana.
Bipincin He is an assistant professor at the Center for Life Sciences at Mahindra University, Hyderabad, Telangana.
Arijit Mytra He is an associate professor in the Faculty of Engineering and Technology, BML Munjal University, Gurugram, Haryana.
NewsLaundry It is a reader-supported, ad-free, independent news outlet based in New Delhi. Support their journalism here.
