Protein Engineering and 3D Printing

Over two decades ago, a groundbreaking achievement led by David Baker at the University of Washington marked the inception of de novo protein design. While ‘Top7,’ an early creation, initially lacked biological functionality, today, the field has evolved into a powerful tool for tailor-made enzymes and proteins. Neil King, collaborating with Baker’s team, emphasizes the transformative nature of this progress, citing the ability to accomplish tasks that were deemed impossible just a year and a half ago. The success owes much to extensive datasets linking protein sequences to structures and the indispensable role of deep learning in unraveling the hidden grammar of protein architecture.

Deep Learning for Protein Design

The utilization of large language models (LLMs), such as those powering ChatGPT, has paved the way for ‘sequence-based’ strategies in protein design. Noelia Ferruz and her team at the Molecular Biology Institute of Barcelona developed ProtGPT2, an algorithm consistently producing synthetic proteins with stable folding. Meanwhile, ‘structure-based’ approaches, fueled by diffusion models akin to those in image-generating tools like DALL-E, demonstrated notable progress in bespoke protein-design algorithms in 2023. RFdiffusion software and the Chroma tool exemplify the power of these approaches in engineering novel proteins.

Deepfake Detection

The surge in generative AI algorithms has facilitated the creation of convincing yet entirely artificial media, raising concerns about potential weaponization. Siwei Lyu at the University at Buffalo highlights the prevalence of AI-generated ‘deepfake’ content, prompting a cat-and-mouse game with media forensics specialists. Solutions include embedding hidden signals in AI models’ output to create watermarks, focusing on content analysis, and the development of detection tools like the DeepFake-O-Meter. As geopolitical conflicts persist, the battle against AI-generated misinformation is anticipated to endure.

CRISPR-Based Gene Editing

The groundbreaking approval of the first-ever CRISPR-based gene-editing therapy in late 2023 marked a significant milestone in genome editing for sickle-cell disease and transfusion-dependent β-thalassaemia. CRISPR and its derivatives, leveraging programmable RNA and DNA-cutting enzymes like Cas9, have revolutionized genetic modifications in the lab. Emerging solutions, such as single-stranded annealing proteins (SSAPs) and prime editing, offer possibilities for precise and programmable insertion of DNA sequences, opening avenues for mitigating the effects of genetic diseases.

Innovations in Crop Engineering

Beyond human health, CRISPR-based technologies are making strides in crop engineering. Caixia Gao and her team at the Chinese Academy of Sciences developed PrimeRoot, a method utilizing prime editing to introduce specific target sites for the insertion of DNA in crops like rice and maize. The technology holds promise for imparting disease and pathogen resistance to crops, showcasing ongoing innovation in CRISPR-based plant genome engineering with broad applicability across plant species.

By Impact Lab