Dr. Berger studies treatments for eye conditions such as neovascular age-related macular degeneration (nAMD) and diabetic macular edema (DME), focusing on novel therapies to improve patient outcomes. He also investigates how dietary fibers affect gut health and metabolic responses, especially in individuals with excess body weight. Additionally, his work examines responses to infections, such as tuberculosis, and the effects of hormone use on conditions like central serous chorioretinopathy, helping to refine treatment strategies for these ailments.
Key findings
In a trial involving 56 patients with nAMD and DME, the new treatment DS-7080a did not significantly improve vision or reduce retinal swelling compared to ranibizumab.
In a study of dietary fibers, acacia gum reduced the hunger hormone ghrelin and improved gut health symptoms over six weeks among participants with excess body weight.
Patients using exogenous hormones were found to have a 5.2 times higher risk of recurrence of central serous chorioretinopathy after treatment, suggesting a need for better hormone management.
Frequently asked questions
Does Dr. Berger study eye conditions?
Yes, Dr. Berger researches treatments for conditions like neovascular age-related macular degeneration and diabetic macular edema.
What treatments has Dr. Berger researched?
He has explored therapies involving monoclonal antibodies and dietary fibers for improving health outcomes in various conditions.
Is Dr. Berger's work relevant to patients with tuberculosis?
Yes, his research on lung pathology in tuberculosis may lead to new treatment strategies for this serious infection.
Can Dr. Berger's studies help those experiencing anxiety?
Indirectly, his research into brain circuits could enhance understanding of decision-making under anxiety, benefiting future treatment approaches.
What is the focus of Dr. Berger's research on dietary fibers?
He studies how different types of dietary fibers affect gut bacteria and health outcomes, especially in individuals who are overweight.
Publications in plain English
Effective Management of a Rare Case of Pediatric ANCA-Associated Vasculitis With Rituximab and Mycophenolate Mofetil.
2026
The American journal of case reports
Fang Y, Kurady A, Hashim FQ, Berger B, Rao A
Plain English This study looks at a rare childhood illness called ANCA-associated vasculitis (AAV), which can cause serious kidney problems. It reports on a 6-year-old girl who was treated with rituximab and mycophenolate mofetil after her initial treatment with steroids failed. Thanks to this new treatment, she showed significant improvement and remained healthy at a 5-year follow-up, demonstrating that these medications can effectively manage AAV in children.
Who this helps: This benefits pediatric patients with ANCA-associated vasculitis and their doctors.
Single-cell and spatial profiling highlights TB-induced myofibroblasts as drivers of lung pathology.
2026
The Journal of experimental medicine
Mbano IM, Liu N, Wadsworth MH, Chambers MJ, Mpotje T +25 more
Plain English This study looked at how tuberculosis (TB) affects lung cells and the damage it causes in the body. Researchers found 30 different types of cells in lung tissues from both TB-infected and healthy individuals, with a particular focus on a specific type of fibroblast that was linked to more severe disease and higher levels of bacteria. Understanding these cells could lead to new treatments for TB, which currently causes about 1.3 million deaths each year.
Who this helps: This benefits patients suffering from tuberculosis and researchers developing new treatments.
Learning the language of protein-protein interactions.
2026
Nature communications
Ullanat V, Jing B, Sledzieski S, Berger B
Plain English This study focused on a new tool called MINT, designed to better understand how proteins work together, which is important for studying diseases. Researchers trained MINT on a large set of data about protein interactions and found that it outperformed other models in predicting how proteins bind and how mutations affect these interactions. For example, MINT provided accurate predictions for how antibodies could neutralize different strains of SARS-CoV-2.
Who this helps: This helps researchers and doctors working on treatments for diseases.
Contribution of terrestrial processes in reducing environmental mycotoxin levels: a review on mycotoxin-soil interactions.
2026
Mycotoxin research
Muñoz K, Korz S, Meyer M, Berger B
Plain English This study looked at how different natural processes in the soil can help reduce harmful mycotoxin levels, which are toxic substances produced by fungi that can affect both people and animals. It found that the main way mycotoxins break down in the soil is through microbial activity, particularly when bacteria are abundant. Additionally, managing nitrogen and pesticide use can help lower fungal infestations, which produce mycotoxins, highlighting the importance of healthy soil for reducing these toxins.
Who this helps: This benefits farmers and agricultural scientists who want to ensure safer crops.
Profile of David Baker, Demis Hassabis, and John Jumper: 2024 Nobel laureates in chemistry.
2026
Proceedings of the National Academy of Sciences of the United States of America
Berger B, Eisenberg DS
Plain English This research focuses on two main achievements in understanding proteins, which are crucial for many biological processes. First, David Baker's lab developed techniques to design proteins based on their amino acid sequences, while Demis Hassabis and John Jumper used artificial intelligence to predict protein structures accurately. This is significant because it could lead to advancements in drug design and understanding diseases better, with the research marking a major step after 40 years of work in this field.
Who this helps: This benefits patients, researchers, and healthcare providers by improving treatments and therapies.
Cortical-limbic circuit dynamics of approach-avoidance conflict in humans.
2026
Nature communications
Staveland BR, Oberschulte J, Berger B, Minarik T, Kim-McManus O +8 more
Plain English This study looked at how the brain decides whether to approach something positive or avoid something negative, particularly in relation to anxiety. Researchers observed brain activity in 20 patients with epilepsy while they played a game resembling Pac-Man, where they had to balance earning points against the risk of losing them to "ghosts." They found that when patients chose to approach rewards, specific brain areas showed increased activity and connectivity, while avoiding threats led to decreased activity, highlighting how these brain circuits influence decision-making in anxious situations.
Who this helps: This benefits patients with anxiety disorders and their doctors by improving understanding of brain processes involved in avoidance behavior.
OzBarley: A genetic and phenotypic data resource capturing the Australian barley breeding history.
2026
Scientific data
Baumann U, Kalashyan E, Schwerdt J, Box A, Brien C +16 more
Plain English The study created OzBarley, a detailed resource that combines genetic and physical data from barley plants grown in Australia to improve barley breeding. It includes data from both older and newer barley types, and tools for measuring important traits like plant growth and seed size. This resource aims to help researchers find the genetic factors that make barley more resilient and productive, which is crucial for food security and sustainable agriculture.
Who this helps: This benefits researchers, farmers, and the agricultural industry as a whole.
Constrained Diffusion as a Paradigm for Evolution.
2026
bioRxiv : the preprint server for biology
Lazarev D, Sappington A, Chau G, Zhang R, Berger B
Plain English This study examined how evolution works by modeling it as a process affected by various biological and environmental factors. Researchers developed a tool called DiffEvol that analyzes genetic data, particularly from the SARS-CoV-2 virus, to understand constraints on its evolution. They found that, following the introduction of the COVID-19 vaccine, the virus experienced significant changes, which they could track more clearly than before. This approach can help predict how viruses evolve and may improve strategies to combat emerging strains.
Who this helps: This benefits researchers studying viruses, healthcare professionals, and public health officials.
Evolutionary dynamics under phenotypic uncertainty.
2026
bioRxiv : the preprint server for biology
Mohanty V, Sappington A, Shakhnovich EI, Berger B
Plain English This research focused on understanding how uncertainty in physical traits (phenotypes) affects evolution. The scientists found that traditional models of population genetics, which have been used for decades, miss key factors, leading to incorrect predictions about how species evolve under uncertain conditions. Their new model shows that high-fitness traits can speed up adaptations, even with few genetic changes, and can help explain why certain low-fitness traits can exist alongside dominant successful traits.
Who this helps: This benefits researchers and doctors studying cancer and bacterial infections, as it can lead to better strategies for treatment.
AI-based methods for simulating, sampling, and predicting protein ensembles.
2026
Current opinion in structural biology
Jing B, Berger B, Jaakkola T
Plain English This study looks at new AI methods that help predict and simulate groups of protein structures, known as protein ensembles. The researchers reviewed various techniques, including coarse-grained models and generative methods, to assess their strengths and weaknesses. They emphasize the importance of improving how these models learn and make predictions, which is crucial for advancing our understanding of proteins.
Who this helps: This benefits researchers and scientists working in drug discovery and protein-related diseases.
Structure and Dynamics of Paramagnetic Iron(II) Sites from the DFT-Assisted Solid-State NMR of Molecular and Surface Species.
2026
Journal of the American Chemical Society
Balavoine G, Carvalho JP, Sanders KJ, Kervern G, Berger B +4 more
Plain English This study looked at the structure and behavior of iron sites in catalysts, particularly those embedded in silica, which are important for processes like converting methane. Researchers used advanced techniques to find that these iron sites are very different when they are on the surface of silica compared to when they are in molecular form. Specifically, they discovered that the surface sites are more uniform and dynamic, which could improve their efficiency in catalysis.
Who this helps: This benefits researchers and engineers working on developing better catalysts for industrial processes.
Assessing the Safety and Efficacy of Agonistic Monoclonal Antibody against Robo4 versus Ranibizumab in Neovascular Age-Related Macular Degeneration and Diabetic Macular Edema: A Phase I, Open-Label, Multicenter Study.
2026
Ophthalmology science
Afridi R, Senaldi G, Hwang JJ, Nguyen N, Inoue T +5 more
Plain English This study looked at a new treatment called DS-7080a, a type of antibody, for patients with eye conditions like neovascular age-related macular degeneration (nAMD) and diabetic macular edema (DME). It involved 56 patients and found that, while DS-7080a was mostly safe, it did not improve vision or reduce retinal swelling compared to the existing treatment ranibizumab. Specifically, six patients experienced side effects, though none were serious, indicating that the new treatment might not be better than what's already available.
Who this helps: This research primarily benefits patients with nAMD and DME by informing them about treatment options.
CANBERRA: A Phase II Randomized Clinical Trial to Test the Therapeutic Potential of Oral Vicasinabin in Diabetic Retinopathy.
2025
Ophthalmology science
Armendariz BG, Luhman UFO, Berger B, Hernandez-Sanchez J, Bogman K +5 more
Plain English The CANBERRA study examined whether an oral medication called vicasinabin could help improve diabetic retinopathy, a serious eye condition affecting people with diabetes that can cause blindness. Researchers found that after 36 weeks, only a small percentage of patients showed improvement in their condition: 7.9% for the placebo group, 9.5% for the 30 mg dose, and 5.7% for the 200 mg dose, with no significant differences between the groups. While vicasinabin was generally safe to use without major side effects, it did not have the expected positive effect on eye health, which highlights the need for better treatments.
Who this helps: This study informs doctors and researchers seeking effective treatments for diabetic retinopathy in diabetic patients.
Circuit dynamics of approach-avoidance conflict in humans.
2025
bioRxiv : the preprint server for biology
Staveland BR, Oberschulte J, Berger B, Minarik T, Kim-McManus O +6 more
Plain English This study examined how people make decisions when faced with choices that involve both rewards and potential dangers, using brain activity measurements from epilepsy patients. Researchers found that when patients chose to approach rewards, certain brain regions showed increased activity, but this dropped significantly when they decided to avoid threats. This understanding of how the brain processes approach and avoidance can help in developing better treatments for anxiety disorders, potentially improving decision-making and reducing excessive avoidance in patients.
Who this helps: This research benefits patients with anxiety disorders.
Chickpea displays a temporal growth response to Mesorhizobium strains under well-watered and drought conditions.
2025
Physiologia plantarum
Iqbal N, Brien C, Jewell N, Berger B, Zhou Y +2 more
Plain English This study looked at how different strains of Mesorhizobium bacteria affect the growth of chickpeas under both normal and drought conditions. Researchers found that one strain, CC1192, helped chickpeas grow better when there was enough water, but both strains performed similarly during drought. There were significant variations in plant growth, with a 3-fold difference in biomass and a nearly 4-fold difference in the number of nodules per plant, which shows that the choice of bacteria can greatly influence crop success depending on the environment.
Who this helps: This benefits farmers and agricultural researchers focused on improving chickpea yields in varying weather conditions.
Becoming Bedridden and Being Bedridden: Implications for Nursing and Care for Older People in Long-Term Care: A Scoping Review.
2025
International journal of older people nursing
Berger B, Graeb F, Baumann M, Wolke R
Plain English This study looked at how older people, especially those in nursing homes, become bedridden and what that means for their care. Researchers reviewed 250 articles and focused in detail on 47, finding that most research centers on the reasons people become bedridden, but very few discuss how to help those individuals maintain some independence while in bed. Understanding these issues is important because it can lead to better care strategies and support for the aging population.
Who this helps: This helps patients in nursing homes and their caregivers.
Secure and federated genome-wide association studies for biobank-scale datasets.
2025
Nature genetics
Cho H, Froelicher D, Chen J, Edupalli M, Pyrgelis A +3 more
Plain English The study focused on improving how researchers can work together to analyze genetic data from large biobanks without compromising privacy. It introduced a new method called secure federated genome-wide association studies (SF-GWAS), which allows data from different institutions to be shared and analyzed efficiently while keeping individual information secure. This approach was tested on several datasets, including one with 410,000 people, and showed a significant increase in speed—up to ten times faster than older methods.
Who this helps: This benefits researchers and institutions working on genetic research and patient care by allowing them to collaborate more effectively while protecting privacy.
Rapid and accurate prediction of protein homo-oligomer symmetry using Seq2Symm.
2025
Nature communications
Kshirsagar M, Meller A, Humphreys IR, Sledzieski S, Xu Y +7 more
Plain English This study developed a new tool called Seq2Symm that quickly and accurately predicts how proteins made of identical subunits (homo-oligomers) are structured. The model showed significant improvements, achieving an average score of around 0.49 for predicting protein symmetry, compared to earlier methods that scored only about 0.25. This is important because knowing the structure of proteins helps in understanding their functions and creating new treatments.
Who this helps: This helps researchers and drug developers in biology and medicine.
ralphi: a deep reinforcement learning framework for haplotype assembly.
2025
bioRxiv : the preprint server for biology
Battistella E, Maheshwari A, Ekim B, Berger B, Popic V
Plain English This study looks at a method called "haplotype assembly," which is about figuring out the specific combinations of genes inherited from our parents. The researchers developed a new tool called ralphi that uses advanced computer techniques to improve the accuracy of this process. They found that ralphi has lower error rates when assembling haplotypes compared to existing methods, achieving better results especially with long gene sequences, reinforcing its effectiveness across commonly tested human genomes.
Who this helps: This helps researchers and doctors who work on genetics and personalized medicine.
Learning the language of protein-protein interactions.
2025
bioRxiv : the preprint server for biology
Ullanat V, Jing B, Sledzieski S, Berger B
Plain English Researchers developed a new tool called MINT to better understand how proteins interact with each other, which is important for studying diseases and cellular functions. MINT outperformed previous models in predicting how strongly proteins bind and how mutations affect these interactions, accurately relating its findings to experimental data. This tool is especially useful for studying antibody responses to variants of COVID-19 and can provide insights into cancer-related protein interactions.
Who this helps: Patients, doctors, and researchers in biomedical fields.
Tracing the Shared Foundations of Gene Expression and Chromatin Structure.
2025
bioRxiv : the preprint server for biology
Liang H, Berger B, Singh R
Plain English This study looked at how the three-dimensional structure of chromatin, called topologically associating domains (TADs), affects how genes are expressed in humans and mice. Researchers found that TADs help regulate genes by bringing them closer together, especially during early development, and that their effectiveness decreases as organisms age. They also discovered that cancer cells change how they use TADs when undergoing chemotherapy, which may influence how those cells adapt.
Who this helps: This helps patients undergoing cancer treatment and researchers studying gene regulation.
Longitudinal profiling of low-abundance strains in microbiomes with ChronoStrain.
2025
Nature microbiology
Kim Y, Worby CJ, Acharya S, van Dijk LR, Alfonsetti D +8 more
Plain English This research developed a new tool called ChronoStrain to help scientists better track and study low-level bacteria in our microbiomes over time. The study found that ChronoStrain was more accurate than previous methods, especially for identifying harmful strains of bacteria, such as Escherichia coli in women with urinary tract infections and Enterococcus faecalis in infants. This is important because accurately detecting these low-abundance bacteria can lead to better treatments and understanding of infections.
Who this helps: Patients with recurring infections, particularly women and infants.
The orexin 1-selective receptor antagonist nivasorexant is a time-dependent inhibitor of CYP2C19 and CYP3A4.
2025
Xenobiotica; the fate of foreign compounds in biological systems
Treiber A, Drouet F, Seeland S, Willecke F, Williams JT +3 more
Plain English This study looked at the drug nivasorexant, which is being tested as a potential treatment for eating disorders, and how it interacts with specific liver enzymes (CYP2C19, CYP3A4, and CYP2C9) that are important for drug metabolism. The researchers found that nivasorexant and its metabolites strongly inhibit CYP2C19 and also affect CYP3A4 in complex ways that weren't expected based on standard tests. These findings are important because they suggest that patients taking nivasorexant may experience unexpected effects from other medications due to these interactions.
Who this helps: This helps patients who are considering nivasorexant for eating disorders, especially those on other medications.
C-H Functionalization via Single Atom Metathesis of C-H and C-X Bonds.
2025
Journal of the American Chemical Society
Jankins TC, Berger B, Aouane FA, Barbeira-Arán S, Didier C +3 more
Plain English This research studied a new way to modify chemical bonds in organic compounds, focusing on the challenging task of changing C-H bonds into halogenated forms. The scientists developed a method that allows for swapping hydrogen atoms and halogens between molecules, making it possible to selectively add or remove these elements. This method can help reduce harmful pollutants by utilizing existing halogenated compounds, making processes in synthetic chemistry more efficient.
Who this helps: This helps chemists and the pharmaceutical industry by providing new methods for creating useful compounds.
Inadvertent intrathecal application of vindesine and its neurological outcome: case report and systematic review of the literature.
2025
Brain & spine
Babaee A, Yasin H, Berger B, Simon M
Plain English This study looked at the dangerous effects of mistakenly injecting a cancer treatment drug, vindesine, into the spinal area instead of a vein, which can lead to severe neurological damage. Among 31 cases reviewed, only 6 out of 31 patients survived this accidental injection, with patients who received a specific treatment involving rinsing the spinal fluid showing a better survival rate of 40%. However, all survivors still experienced significant lasting neurological issues.
Who this helps: This information is vital for doctors treating cancer patients, particularly in how to manage such dangerous mistakes.
Predictors of fatigue improvement in multimodal, multimodal-aerobic and aerobic exercise intervention studies in breast cancer survivors with cancer-related fatigue.
2025
Scientific reports
Kröz M, Reif M, Fässler-Teal LR, Berger B, Sasselli C +4 more
Plain English This study looked at how certain factors might help breast cancer survivors feel less tired after treatments for cancer-related fatigue (CRF). It found that two specific traits—autonomic regulation and internal coherence—significantly predicted improvements in fatigue even four years after treatment, with strong statistical support (p<0.001). Understanding these predictors is important because it can help tailor therapies to better support survivors in managing their fatigue.
Who this helps: This helps breast cancer survivors dealing with fatigue after treatment.
The phenomenon of bedriddenness in older adults in hospitals: A scoping review of current research and nursing implications.
2025
Geriatric nursing (New York, N.Y.)
Berger B, Graeb F, Baumann M, Wolke R
Plain English The research looked at the issue of older adults becoming bedridden in hospitals, a problem that's becoming more common as the population ages. Out of 55 studies reviewed, most focused on the negative effects of being bedridden but fewer on how to prevent it; the study highlighted that even short periods of staying immobile can harm patients. It stressed the need for better strategies to encourage movement, such as exercise programs, to help older adults maintain their independence.
Who this helps: This benefits older patients in hospitals and the healthcare workers caring for them.
Correction: Predictors of fatigue improvement in multimodal, multimodal-aerobic and aerobic exercise intervention studies in breast cancer survivors with cancer-related fatigue.
2025
Scientific reports
Kröz M, Reif M, Fässler-Teal LR, Berger B, Sasselli C +4 more
Sparse autoencoders uncover biologically interpretable features in protein language model representations.
2025
Proceedings of the National Academy of Sciences of the United States of America
Gujral O, Bafna M, Alm E, Berger B
Plain English This study looked at how to make protein language models (PLMs), which help predict protein structures and their functions, easier to understand. Using a method called sparse autoencoders, the researchers were able to pull out clear features related to proteins, finding that many of these features matched known biological terms. They found that these features were better at revealing useful information about proteins than traditional methods, which could make it easier to trust and fully utilize these AI tools in biology.
Who this helps: This research benefits scientists and researchers working with protein data and AI technologies.
[Bedriddenness and sociospatial participation-A reflection].
2025
Zeitschrift fur Gerontologie und Geriatrie
Berger B, Schnabel M
Plain English This study looked at how being bedridden affects older people's ability to participate in social activities, especially in nursing homes. Researchers found that the way residents and their surroundings are organized greatly influences their opportunities for interaction. They emphasize that better design of these environments is crucial to enhance the social participation of bedridden individuals.
Who this helps: This helps patients who are bedridden, particularly older adults in nursing homes.
Fasting elicits gut microbiome signature changes that extend to type 1 diabetes patients.
2025
Frontiers in endocrinology
Graef FA, Berger B, Bahr LS, Stange R, Michalsen A +3 more
Plain English This study looked at how prolonged fasting affects the gut microbiome in patients with type 1 diabetes (T1D). Researchers found that after fasting, the gut microbiome of T1D patients changed significantly, becoming similar to that of non-diabetic individuals. This was important because the changes in gut bacteria were also linked to improvements in cholesterol and blood pressure in T1D patients, indicating that fasting may help manage these conditions.
Who this helps: This benefits patients with type 1 diabetes and healthcare providers managing their treatment.
Unveiling causal regulatory mechanisms through cell-state parallax.
2025
Nature communications
Wu AP, Singh R, Walsh CA, Berger B
Plain English This research studied how specific genetic variations linked to diseases, particularly schizophrenia, affect gene activity in brain cells. Using a new tool called GrID-Net, the researchers found that these variations can disrupt the normal function of 132 genes, including important ones for potassium transport. Their approach improved the understanding of how noncoding DNA affects gene expression by 36%, which is significant for developing targeted therapies.
Who this helps: This helps patients with schizophrenia and researchers working to develop new treatments.
Learning ecosystem-scale dynamics from microbiome data with MDSINE2.
2025
Nature microbiology
Gibson TE, Kim Y, Acharya S, Kaplan DE, DiBenedetto N +5 more
Plain English This research focused on enhancing the understanding of microbial ecosystems, particularly the gut microbiome, using a new tool called MDSINE2. The study demonstrated that MDSINE2 can effectively analyze complex microbiome data, outperforming previous methods by identifying specific groups of microbes that interact with one another in predictable ways. This is significant because it helps scientists better understand how these microbial communities respond to changes, such as diet or antibiotics, which could lead to better approaches to manage gut health.
Who this helps: This benefits patients with gut health issues and researchers studying the microbiome.
Learning residue level protein dynamics with multiscale Gaussians.
2025
ArXiv
Bafna M, Jing B, Berger B
Plain English Researchers developed a new tool called DynaProt to better understand how proteins move and flex based on their static structures. This tool can predict details about protein dynamics with impressive accuracy, requiring significantly fewer resources than traditional methods—specifically, it can estimate how flexible each part of the protein is and relationships between different parts. This is important because it can help scientists and doctors understand protein functions and diseases more efficiently.
Who this helps: Patients and doctors studying diseases related to protein dysfunction.
Generating functional and multistate proteins with a multimodal diffusion transformer.
2025
bioRxiv : the preprint server for biology
Jing B, Sappington A, Bafna M, Shah R, Tang A +4 more
Plain English This study focused on creating new proteins using a method called ProDiT, which combines different ways of modeling protein sequences and shapes. Researchers worked with data from 214 million proteins and were able to design unique proteins that maintain important functional features and can perform a variety of tasks—specifically across 465 different functions. This matters because it opens up possibilities for developing proteins with specific abilities that current methods cannot easily achieve.
Who this helps: This benefits researchers and scientists working in biotechnology and medicine.
On the Viability of Carbonyl Hydroboration Catalysed by Aluminium Hydrides.
2025
Angewandte Chemie (International ed. in English)
Caise A, Berger B, Murray A, Kolychev E, Hicks J +4 more
Plain English This study looked at a chemical process called hydroboration, specifically how certain catalysts made with aluminum help in transforming compounds known as aldehydes and ketones. The researchers found that the previously accepted mechanisms didn't actually occur, and in fact, an experiment with a compound called acetophenone showed no conversion at all when using the proposed catalyst, despite earlier claims of 51% conversion. This is important because it challenges established ideas about how these reactions work and suggests that previous findings might not be reliable.
Who this helps: This benefits chemists and researchers working with catalysts in organic chemistry.
Phenotyping of industrial hemp () genotypes with different growth habits.
2025
Functional plant biology : FPB
Gill AR, Loveys BR, Brien CJ, Jewell N, Berger B +2 more
Plain English This study looked at different types of industrial hemp plants to see how they grow and respond to dry conditions. Researchers found that all three hemp varieties, despite their different shapes, showed similar declines in growth when water was scarce; for instance, all types had lower plant height and fewer leaves when water was limited. The short, bushy hemp variety, Han NE, might use water more efficiently, which could be important for farming practices.
Who this helps: This information benefits farmers and plant breeders by guiding them in choosing the best hemp varieties for challenging growing conditions.
The current role of surgery for single brain metastases.
2025
Brain & spine
Ersoy TF, Brainman D, Dimov D, Coras R, Berger B +2 more
Plain English This study looked at the role of surgery for patients with single brain metastases, which are cancerous growths that spread to the brain. Researchers reviewed data from 202 patients who had surgery between 2015 and 2023 and found that removing all of the tumor did not lead to better survival or progression-free rates; 69.3% had complete removal, but that didn't improve outcomes significantly. The study emphasizes that while fewer complications were seen in asymptomatic patients, having small leftover tumors does not greatly affect survival, highlighting the growing importance of surgery to gather tissue for treatment insights.
Who this helps: This helps patients with brain metastases and their doctors by refining treatment approaches.
Memory-efficient, accelerated protein interaction inference with blocked, multi-GPU D-SCRIPT.
2025
Bioinformatics (Oxford, England)
Schäffer DE, Sledzieski S, Cowen L, Berger B
Plain English This study focuses on improving a tool called D-SCRIPT, which is used to identify how proteins interact with each other. The researchers found that a new method using multiple graphics processing units (GPUs) significantly cuts down the memory needed for these calculations, making the process 13.8 times more efficient for large sets of proteins. This is important because it allows scientists to analyze protein interactions more quickly and with fewer resources, which can speed up research in various fields, including drug development.
Who this helps: This helps researchers and scientists studying proteins in areas like medicine and biotechnology.
Impact of discontinuing automatic reflex urine culture after urinalysis: a diagnostic and antibiotic stewardship initiative.
2025
Frontiers in cellular and infection microbiology
Berger B, Lukey J, Jinadatha C, Navarathna DH
Plain English This study looked at what happened when the Central Texas Veterans’ HealthCare System stopped automatically testing urine samples for infections unless doctors specifically requested it. After this change, the number of urine cultures processed dropped significantly, leading to a 40% reduction in the use of the broad antibiotic ciprofloxacin and an increase in the more targeted antibiotic nitrofurantoin. This matters because it means less unnecessary antibiotic use, which helps fight antibiotic resistance and allows labs to focus on more relevant cases of urinary tract infections.
Who this helps: This benefits patients by ensuring they receive appropriate antibiotics for their infections.
Cost patterns of long-term prescribed medications in patients with multiple sclerosis in Germany.
2025
Scientific reports
Hecker M, Berger B, Frahm N, Heidler F, Barhoum B +4 more
Plain English This study examined the costs of long-term medications for 728 patients with multiple sclerosis (MS) in Germany. It found that the average annual cost of medication was €11,788, with 92.1% of this cost coming from disease-modifying drugs (DMDs). Younger patients with relapsing MS had higher costs due to more frequent use of these expensive medications, highlighting the need for effective early treatment to reduce future expenses.
Who this helps: This information benefits patients with MS and the healthcare providers who care for them.
Hyperspectral-based classification of individual wheat plants into fine-scale reproductive stages.
2025
Plant methods
Xie Y, Roy SJ, Schilling RK, Berger B, Liu H
Plain English This study looked at how to better classify individual wheat plants at different growth stages using advanced imaging technology. Researchers found that by using a method called Support Vector Machine classification, they achieved a high accuracy rate of 83.2% for identifying plants before they flower by combining various imaging techniques. This is important because it can reduce the heavy reliance on manual inspections, making the process cheaper and more efficient while also enhancing safety during tests of genetically modified plants.
Who this helps: This helps farmers, researchers, and agricultural scientists involved in biotechnology and crop management.
Graph-based deep reinforcement learning for haplotype assembly with Ralphi.
2025
Genome research
Battistella E, Maheshwari A, Ekim B, Berger B, Popic V
Plain English This research focused on improving how scientists piece together genetic information from maternal and paternal chromosome copies, a process known as haplotype assembly. The new method, called Ralphi, uses advanced computer techniques to more accurately organize DNA reads into their respective haplotypes, resulting in lower error rates. Specifically, Ralphi outperformed existing methods by achieving better accuracy over longer segments of genetic data while using both short and long DNA reads in tests.
Who this helps: This benefits genetic researchers and healthcare providers by improving the understanding of genetic variations in patients.
Exogenous Hormone Use as a Predictor of Central Serous Chorioretinopathy Recurrence: A Comparative Analysis of Observation Versus Low-Fluence Photodynamic Therapy.
2025
Journal of vitreoretinal diseases
Varghese D, Campagna G, Elwood K, Timmons K, Jhaveri C +3 more
Plain English This study looked at patients with central serous chorioretinopathy (CSCR), a condition that often clears up on its own but can come back. Researchers found that among 90 patients, those who used exogenous hormones were more likely to have recurrences after treatment, with a significant increase in recurrence risk (5.2 times higher) in hormone users. This is important because it suggests that managing hormone use could help lower the chances of CSCR coming back.
Who this helps: This helps patients at risk for CSCR, particularly those using hormones.
Mechanistic insights into microbiome-dependent and personalized responses to dietary fibre in a randomized controlled trial.
2025
medRxiv : the preprint server for health sciences
Armet AM, Li F, Deehan EC, Nikolaeva DD, Delannoy-Bruno O +23 more
Plain English This study looked at how different types of dietary fiber affect people with excess body weight and found that two types, acacia gum and resistant starch type 4, changed gut bacteria and fatty acid levels, while improving inflammation and feelings of fullness—regardless of fiber fermentation. Over six weeks, acacia gum specifically reduced ghrelin, a hormone that increases hunger, and predictions showed potential fiber-specific impacts on blood pressure and inflammation markers from microbial data. These insights help create tailored dietary fiber recommendations, making fiber supplementation more effective for individual needs.
Who this helps: Patients with excess body weight looking to improve their health.
Shechi: A Secure Distributed Computation Compiler Based on Multiparty Homomorphic Encryption.
2025
Proceedings. UNIX Security Symposium
Smajlović H, Froelicher D, Shajii A, Berger B, Cho H +1 more
Plain English This study introduces Shechi, a new programming tool that makes it easier to securely analyze large sets of data spread across different locations. The researchers found that Shechi can improve computation speed by up to 15 times compared to current best methods and can make coding tasks 40 times easier than what expert developers have done manually. This matters because it allows for safer and more efficient processing of sensitive data, such as medical or genetic information, without compromising privacy.
Who this helps: Patients and researchers who handle sensitive information.