Scientists have found that Artificial Intelligence can ‘predict’ the biological language of cancer and neurodegenerative disorders like Alzheimer’s.
Scientists have found that powerful algorithms that can predict the biological language of cancer and Alzheimer’s. These algorithms are used by Facebook, Netflix, and Amazon.
Big data from decades of study have been fed into a machine language model. This determines whether artificial intelligence could make more advanced findings than humans.
University St. John’s College, Cambridge found that the ‘biological language’ of cancer and Alzheimer’s, and other neurodegenerative diseases. This could be deciphered by machine-learning technology.
They published their groundbreaking research on 8 April 2021 in the scientific journal PNAS. This could be used “to correct grammatical errors within cells causing disease” in the future.
Fluorescence Microscopy Condensates in living cells
Fluorescence microscopy of protein condensates in living cells. Credit: Harvard University, Weitz laboratory.
Professor Tuomas Knowles, lead author of the paper and Fellow at St. John’s College, said. “The introduction of machine learning to research on neurodegenerative diseases and cancer is an utter transformer. The ultimate goal is to use artificial intelligence to create targeted medicines to significantly alleviate symptoms or to avoid dementia.
Any time Netflix recommends a series or Facebook suggests that someone is friendly. Platforms make a highly informed guess about the next thing people are using powerful machine-learning algorithms. Voice helpers like Alexa and Siri will also identify individuals and speak to you immediately.
Dr. Kadi Liis Saar, the first author of the article, and a St John’s College Researcher used similar machine-learning technology. He used them to learn a wide-ranging language model to examine what happens when something is wrong with proteins within the body to cause disease.
She said: ‘Millions and thousands of proteins are found in the human body and many scientists still don’t know the role of many of them. We requested a neural network-based language model to learn protein language.
Condensates forming Fluorescence Microscopy Protein
Fluorescence microscopy of protein condensates in living cells. Credit: Harvard University, Weitz laboratory.
“We explicitly asked the software for learning the language of biomolecular form shifting condensates – goths of cell proteins – that scientists really need to understand to crack the language of biofunction and malfunction that cancer and neurodegenerative diseases such as Alzheimer’s. Without being specifically told, we found that scientists have already discovered over decades of research in protein language.”
Proteins are large, complex molecules which play a large number of critical roles in the body. They work mostly in cells and are needed for the structure, function and control of the tissue and organ of the body – for example, antibodies are a protein that protects the body.
The disorders of Alzheimer’s, Parkinson’s- and Huntington’s are three of the most common neurodegenerative diseases.
Proteins are rogue, they form clumps, and destroy healthy nerve cells in Alzheimer’s disease, which affects 50 million people worldwide. A healthy brain has a quality control mechanism, with these potentially toxic protein masses, known as aggregates, efficiently disposed of.
Researchers believe now that such disordered proteins are often liquid-like goutlets of proteins called condensates that don’t have a membrane and fuse freely. In contrast to irreversible protein aggregates, protein condensates can form and reform and are often compared to shapeshifting wax blobs in lava lamps.
Professor Knowles said: “Condensates of proteins have recently received considerable attention in the scientific world because they regulate key events in cells such as gene expression – how our DNA is translated into proteins – and the synthesis of proteins – how cells produce proteins.
“These protein droplet-related defects could lead to diseases such as cancer. Therefore incorporating natural language processing technology into research into the molecular roots of protein failure is crucial if the grammatical errors in cells that cause disease are to be right.”
Dr. Saar said: “All data on the known proteins have been added to the algorithm so that it can learn and predict the language of proteins as well as these models can learn the human language and how WhatsApp can suggest words to you.”
“Then we were able to inquire about the basic grammar, which only refers to certain proteins forming condensates in cells. It’s a very difficult matter and unlocking it will help us learn the rules of the disease language.”
Due to increasing data availability, increased computational power and technological developments that have produced more efficient algorithms, the machine-learning technology is evolving quickly.
Machine Learning
Further machine-learning could turn future research into cancer and neurodegenerative diseases. Discoveries may go beyond what scientists know and speculate about diseases at present and even beyond what the human brain would comprehend without the assistance of machine learning.
Dr Saar explained: “Machine learning should be free from the constraints of researchers’ opinion that the goals for scientific discovery are new links that we have yet to conceive of. It’s indeed quite exciting.”
The network has now been created for researchers all over the world to make progress by more scientists publicly accessible.
The Entrepreneurs Diaries is now available on Telegram. Join our telegram channel to get instant updates from TED
