A team of microbiologists, chemists and pharmaceutical specialists at Shandong University, Guangzhou Medical University, Second Military Medical University and Qingdao University, all in China, has developed an AI model that generates antimicrobial peptide structures for screening against treatment-resistant microbes.

In their study published in the journal Science Advances, the group developed a compression method to reduce the number of elements needed in training data for an AI system, which helped to reduce diversification issues with current AI models.

Prior research has suggested that drug-resistant microbes are one of the most pressing problems in medical science. Researchers around the world have been looking for new ways to treat people infected with such microbes—one approach involves developing antimicrobial peptides, which work by targeting bacterial membranes.

Unfortunately, developing or finding peptides has proven to be too slow to address the crisis. So researchers have turned to AI-based approaches to aid in finding such peptides. But that approach has encountered problems, as well, the biggest being the lack of a large training base, which leads to peptide discovery that lacks diversity.

In this new study, the researchers in China found a way around this problem by developing a compression technique that reduces the number of elements needed to train their AI system.

The researchers call their system a two-stage AI pipeline leverage diffusion model. The first stage works by compressing data describing 2.8 million known peptides into a numerical form by amplifying signal noise randomly. The second stage then pulls new peptides from the simplified data, removes the noise, and decompresses the data used to describe its peptide sequence.

In testing their new system, the research team found that it was able to filter peptides listed in a training database down to a reasonable number of those most likely to have antimicrobial properties. In looking at 600,000 of them, the team experimentally tested 40 peptides and found 25 that showed promise in combating bacterial and fungal pathogens.

A team of Chinese scientists has used targeted gene editing to develop rice that produces coenzyme Q10 (CoQ10), a vital compound for human health.

Led by Prof. Chen Xiaoya from the CAS Center for Excellence in Molecular Plant Sciences/Shanghai Chenshan Research Center and Prof. Gao Caixia from the Institute of Genetics and Developmental Biology of the Chinese Academy of Sciences (CAS), the researchers used targeted gene editing to modify just five amino acids of the Coq1 rice enzyme, creating new rice varieties capable of synthesizing CoQ10.

The study is published in Cell.

CoQ10 is important for human health, especially heart protection. It is an essential component of the mitochondrial electron transport chain and serves as a fat-soluble antioxidant. Different species synthesize CoQ with different side-chain lengths.

Humans synthesize CoQ10 with a side chain of 10 isoprene units (C50), while cereal food crops like rice and wheat, and certain vegetables and fruits, primarily synthesize CoQ9, which contains nine isoprene units (C45).

The invention of CoQ10-producing crops to substantially increase the CoQ10 content in plant-based foods offers a cost-effective and environmentally friendly approach to nutritional fortification, offering great potential benefit.

The molecular mechanism underlying the variation in CoQ side-chain length was previously unclear. Thanks to the diverse plant collections at the Shanghai Chenshan Botanical Garden of the CAS Center for Excellence in Molecular Plant Sciences, the research team obtained 134 plant samples from 67 families, including mosses, clubmosses, ferns, gymnosperms, and angiosperms.

The researchers determined the distribution patterns of CoQ types in these species and revealed that CoQ10 is an ancestral trait of flowering plants, with most plants still synthesizing CoQ10. However, grasses, daisies, and cucurbit plants mainly produce CoQ9.

After analyzing the evolutionary trajectories and natural variations of Coq1 enzymes across more than 1,000 terrestrial plant species, as well as by using machine learning, the researchers ultimately identified five amino acid sites that determine chain length.

After the targeted editing, the rice plants in this study primarily synthesized CoQ10, with up to 5 μg/g produced per rice grain. This shows that gene editing has become an efficient and safe technology for crop breeding.

The successful development of CoQ10 rice will significantly expand the food sources of CoQ10. It also provides an example of using big data and AI for crop breeding.

Researchers have revealed new details about the CRISPR-Cas5-HNH/Cascade complex, a variant of the type I-E CRISPR-Cas system, providing insights into its DNA recognition and cleavage mechanisms.

The study, published in Nature Communications, was conducted by researchers from the Institute of Physics of the Chinese Academy of Sciences.

The CRISPR-Cas5-HNH/Cascade complex, which serves as an immune defense system in prokaryotes, has long been studied for its ability to protect bacteria from invading genetic materials. However, recent research uncovers a novel mechanism in this complex that differs significantly from the well-known type II CRISPR systems like Cas9.

Researchers utilized cryo-electron microscopy to capture the structures of the Cas5-HNH/Cascade complex in both DNA-bound and unbound states. They revealed striking conformational changes. The complex adopts a more compact structure when bound to double-stranded DNA (dsDNA), with the target DNA strand making a pronounced U-turn and interacting with the HNH nuclease domain.

The target strand is cleaved first, followed by the non-target strand, a departure from the cleavage order seen in other CRISPR systems like Cas12a.

The Cas5-HNH domain itself was found to play a crucial role in the nuclease activity of the Cas5-HNH/Cascade complex. Researchers demonstrated that mutations in key residues of the HNH domain, particularly histidine and aspartate, can completely abolish cleavage activity.

This suggests that the HNH domain is not only essential for target DNA recognition but also for the specific cleavage of DNA, further differentiating the mechanism from other systems.

Moreover, researchers observed that the architecture of the Cas5-HNH/Cascade complex is notably different from canonical type I-E systems. For example, the HNH domain was fused to the C-terminus of Cas5, replacing the need for Cas3, a protein commonly involved in DNA degradation in other systems.

This structural innovation pointed to the complex’s ability to cleave DNA independently of the traditional components seen in similar CRISPR systems.

By elucidating the role of the HNH nuclease domain and revealing the structural changes that occur upon DNA binding, this work not only deepens the understanding of CRISPR-Cas5-HNH but also paves the way for refining CRISPR technologies for therapeutic applications.

Picking out individual cells in distorted microscopy images is now as easy as clicking a button.

A new version of Cellpose—the popular tool that maps the boundaries of diverse cells in microscopy images—now works on less-than-perfect pictures that are noisy, blurry, or undersampled.

Many general methods used for segmenting individual cells in microscopy images, including previous versions of Cellpose, have a hard time recognizing cellular boundaries that have been distorted by noise, blurring, or undersampling.

Janelia Group Leaders Carsen Stringer and Marius Pachitariu set out to address this issue with the development of Cellpose3. Unlike previous approaches, which train models to improve the quality of distorted images, Cellpose3 was instead trained to improve the segmentation of distorted images.

The study is published in the journal Nature Methods.

The Cellpose3 restoration algorithm, when applied to distorted images, produces crisp restored images which can then be easily segmented by the original Cellpose segmentation algorithm.

Cellpose3 is also trained on a large, varied collection of images, enabling users to easily use the new method, which is available as a “one-click” button in the Cellpose application, on their own data.

Plastic is the most prevalent marine pollutant, and plastic surfaces are the fastest growing habitat in the ocean. Researchers at the University of Hawai’i (UH) at Mānoa have recently discovered that many species of fungi isolated from Hawai’i’s nearshore environment have the ability to degrade plastic, and some can be conditioned to do it faster.

The work is published in the journal Mycologia.

“Plastic in the environment today is extremely long-lived, and is nearly impossible to degrade using existing technologies,” said Ronja Steinbach, who led this research as a marine biology undergraduate student in the UH Mānoa College of Natural Sciences.

“Our research highlights marine fungi as a promising and largely untapped group to investigate for new ways to recycle and remove plastic from nature. Very few people study fungi in the ocean, and we estimated that fewer than one percent of marine fungi are currently described.”

For consumers, plastics are cheap, strong, and useful, but plastic waste is problematic, because rather than decomposing, it breaks into microplastics when exposed to sunlight, heat, and physical force. Plastics are harmful to marine ecosystems—they can concentrate dangerous chemicals, such as phthalates and bisphenol A; entrap or harm animals; or be ingested and lead to starvation in marine animals due to malnutrition. With the equivalent of about 625,000 garbage trucks of plastic entering the ocean each year, finding ways to degrade these compounds is critical.