The research and development of peptide drugs and the discovery of peptide lead compounds are expensive, time-consuming, with a high risk for failure. The discovery and design of active peptides is a key step of peptide drug development. The traditional method uses experiments to process peptides or modify the structure of existing peptides, then observe in experiments these structures’ pharmacological activity. This approach involves a complicated process, is labor-intensive, costly, and time-consuming. Furthermore, it cannot predict the relevant physical and chemical properties of the candidate peptides.
XtalPi AI Research Center (XARC) provides active peptide generation, prediction, and recommendation solutions that are based on AI, theoretical calculation, and cloud computing, which can effectively improve R&D efficiency and shorten the peptide drug R&D cycle.
The peptide sequence generation module (XARCP Io algorithm) developed by XARC is based on the deep learning algorithm. Combined with a large-scale database of peptide sequences and properties, XARC IO can rapidly generate millions of new peptide sequences and cover the physicochemical properties space of desired peptides. With the assistance of quantum physics and chemistry algorithms or actual experimental data, we use our machine-learning-based peptide function prediction model, XARCP, combined with expert judgement model, to quickly screen and recommend candidate peptide sequences. XARC’s peptide drug design and function prediction services can be used for the rational design, prediction, and refined multi-dimensional analysis of various types of peptide compounds, including antihypertensive peptides, HIV inhibitory peptides, neuropeptides, hemolysis peptides, anti-parasite peptides, anti-cancer peptides, antimicrobial peptides, cell-penetrating peptides, anti-viral peptides, peptide hormones, the blood-brain barrier peptides, anti-angiogenic peptides, and toxic peptides. With feedback from further experimental results, the model can be optimized in future iterations for increased screening efficiency.