New AI models help identify unknown proteins with high accuracy

AI is revolutionising biology by helping scientists uncover hidden proteins that traditional methods struggle to detect.

Researchers have developed two AI models, InstaNovo and InstaNovo+, designed to identify unknown proteins, which could improve disease research and treatment development.

Proteins, the functional components of cells, often differ from their genetic blueprint due to modifications after production.

Such variations can be difficult to analyse using conventional tools. InstaNovo, inspired by OpenAI’s GPT-4, translates mass spectrometry data into amino acid sequences, while InstaNovo+ refines these results using a noise-reduction technique similar to AI image generation.

Together, they outperform standard methods in complex protein sequencing tasks, particularly for challenging targets like human immune proteins.

Scientists believe these models could help explain biological mysteries, such as how stingrays adapt to different water environments or why pancreatic cancer leads to severe muscle wasting.

While the tools are promising, researchers caution that AI-generated results require verification. Nonetheless, AI sequencing is expected to complement traditional database searches, pushing biological research into new frontiers.

For more information on these topics, visit diplomacy.edu.