Category: Machine Learning (ML) News

Researchers at Mass General Brigham have developed a machine learning algorithm, PAMmla, that predicts the properties of over 64 million CRISPR-Cas9 enzymes. This groundbreaking approach, published in Nature, dramatically expands the toolkit for gene editing, potentially reducing off-target effects and improving therapy safety and efficacy.

Current CRISPR-Cas9 technology faces limitations, including the risk of unintended DNA modifications (off-target effects). PAMmla addresses this by predicting which enzymes are most likely to precisely target the desired genes while minimizing off-target activity. Crucially, the scalability of PAMmla distinguishes it from previous efforts, generating far more potential enzymes for researchers to explore.

The algorithm works by predicting the protospacer adjacent motif (PAM), a short DNA sequence that CRISPR enzymes need to recognize and bind to. By identifying novel PAMs, researchers can engineer Cas9 enzymes with enhanced specificity. Initial proof-of-concept experiments in human cells and a mouse model of retinitis pigmentosa demonstrated that PAMmla-predicted enzymes had greater precision than traditional enzymes.

This research represents a significant step forward in gene and cell therapy. By enabling researchers to predict and customize CRISPR enzymes, PAMmla promises to accelerate the development of safer and more effective gene therapies for a wide range of genetic disorders.

Colin Raffel, a prominent machine learning researcher, recently made headlines for his relocation from North Carolina to Toronto. This move, far from a simple geographical shift, represents a significant statement about the future of technological innovation and accessibility. Raffel’s declaration, “Everyone should have a voice in tech,” encapsulates the driving force behind his decision, highlighting a growing concern within the tech community regarding inclusivity and equitable representation.

The vibrant tech scene in Toronto, a city increasingly recognized as a global hub for artificial intelligence and machine learning, undoubtedly played a significant role in Raffel’s choice. However, his statement suggests that the allure of Toronto extends beyond mere career opportunities. He’s clearly seeking an environment that actively fosters diverse perspectives and challenges the often homogeneous landscape of the tech industry. North Carolina, while possessing its own strengths, may not have offered the same level of access to collaborative networks and initiatives dedicated to broadening participation in the field.

Raffel’s commitment to inclusivity likely stems from a deep understanding of the ethical considerations inherent in machine learning. Algorithms, after all, are trained on data, and biased data inevitably leads to biased outcomes. A diverse team, reflecting the varied experiences and perspectives of the communities these algorithms will ultimately impact, is crucial for mitigating these biases and ensuring fairness and equity. His move, therefore, can be interpreted not just as a personal career advancement, but as a deliberate act of contributing to a more just and representative technological future.

The implications of Raffel’s decision extend beyond his individual career. It serves as a call to action for other researchers, companies, and institutions within the tech industry. It underscores the urgent need for proactive measures to attract and retain talent from underrepresented groups, creating environments where everyone feels empowered to contribute their unique skills and perspectives. Only through such concerted efforts can we hope to build a truly inclusive and equitable technological landscape, one where the benefits of technological advancement are shared broadly and fairly. Raffel’s relocation to Toronto, therefore, is more than a simple news item; it’s a powerful symbol of the growing movement towards a more representative and responsible technological future. His actions inspire others to consider their own roles in fostering inclusivity and ensuring that the “voice” in tech truly reflects the diversity of the world it serves.

Via: Source

A team of researchers from Southwest Research Institute (SwRI), The University of Texas at San Antonio (UTSA), and Texas Biomedical Research Institute (Texas Biomed) have developed a machine learning algorithm that identified over two dozen potential treatments for Nipah and Hendra viruses, zoonotic pathogens that can jump from animals to humans. Using SwRI’s Rhodium™ software, the team mapped the protein structure of a related virus, measles, and virtually screened 40 million compounds to identify promising antiviral candidates.

The research, funded by the Department of Defense, leverages machine learning to overcome the challenges of studying these highly pathogenic viruses in high-containment labs. By virtually screening compounds, researchers can quickly prioritize potential treatments, saving significant time and resources. The approach could potentially lead to broad-spectrum therapies for related viruses, including measles, a significant benefit given the high mortality rates associated with Nipah and Hendra infections (40-75% fatality). The findings were presented at the Hendra@30 Henipavirus International Conference.

This innovative approach underscores the power of AI in accelerating drug discovery for challenging infectious diseases, potentially paving the way for faster development of effective treatments for Nipah and Hendra, and related illnesses.