LZLiron Zahavi

Liron Zahavi

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Liron Zahavi has a total of 55 co-authors. Here are the top co-authors by number of shared publications:

  • Eran Segal

    Department of Computer Science and Applied Mathematics, Weizmann Institute of Science, Rehovot, Israel

    7 shared publications
  • Anastasia Godneva

    Department of Computer Science and Applied Mathematics, Weizmann Institute of Science, Rehovot, Israel

    6 shared publications
  • Adina Weinberger

    Department of Computer Science and Applied Mathematics, Weizmann Institute of Science, Rehovot, Israel

    4 shared publications
  • Michal Rein

    Department of Computer Science and Applied Mathematics, Weizmann Institute of Science, Rehovot, Israel

    4 shared publications
  • Saar Shoer

    Department of Computer Science and Applied Mathematics, Weizmann Institute of Science, Rehovot, Israel

    4 shared publications
  • Daniel Barak

    Department of Computer Science and Applied Mathematics, Weizmann Institute of Science, Rehovot, Israel

    3 shared publications

Latest Publications

  • Top Cited Papers

    Diet–microbiome associations in 10,068 individuals from the Human Phenotype Project to guide personalized nutrition

    by Tomer Segev, Daniel Barak, Liron Zahavi, Anastasia Godneva, Michal Rein, David Krongauz, Dorit Samocha-Bonet, Hagai Rossman, Adina Weinberger, Eran Segal

    Diet is a major environmental factor influencing the human gut microbiome. However, the effects of specific foods and dietary patterns on microbial composition, diversity and function is not fully understood, limiting progress toward personalized dietary strategies. Here, leveraging 10,068 participants from the Human Phenotype Project with app-based diet logs and shotgun metagenomics, we predicted diet–microbiome associations at species-level resolution. Diet significantly predicted microbial diversity (richness r = 0.26, Shannon Index r = 0.24), the relative abundance of 669 of 724 species tested (92.4%, false discovery rate Lawsonibacter asaccharolyticus (r = 0.43), yogurt with Streptococcus thermophilus (r = 0.42) and milk with Bifidobacterium species (r = 0.31–0.36). In parallel, broader dietary patterns, especially the degree of food processing, emerged as predictors of microbial diversity and composition. We also show that diet–microbiome associations persist over four years, with 82.5% of species exhibiting significant longitudinal tracking between predicted and observed abundances. Finally, we developed an exploratory analysis for simulating personalized dietary interventions with predicted microbiome shift effects that are associated with improvements in cardiometabolic health. Our findings demonstrate that diet is strongly associated with microbiome composition, diversity and function, and highlight its potential for guiding personalized interventions.


  • Open Access

    Self-supervised learning enables robust microbiome predictions in data-limited and cross-cohort settings

    by Liron Zahavi, Zachary Levine, Anastasia Godneva, Veronika Dubinkina, Raja Dhir, Katherine S. Pollard, Adina Weinberger, Eran Segal

    The gut microbiome plays a crucial role in human health, but machine learning applications in this field face significant challenges, including limited labeled data availability, high dimensionality, and batch effects across different cohorts. To address these limitations, we developed representation learning models for gut microbiome metagenomic data, drawing inspiration from foundation models approaches based on self-supervised and transfer learning principles. By leveraging a large collection of 85,364 metagenomic samples, we implemented multiple self-supervised learning methods, including masked autoencoders with varying masking rates and adapted single-cell RNA sequencing models (scVI and scGPT), to generate embeddings from bacterial abundance profiles. These learned representations demonstrated significant advantages over raw bacterial abundances in two key scenarios: first, when training predictive models with very limited labeled data, improving prediction performance for age (r = 0.14 vs. 0.06), BMI (r = 0.16 vs. 0.11), visceral fat mass (r = 0.25 vs. 0.18), and drug usage classification (PR-AUC = 0.81 vs. 0.73); and second, when generalizing predictions across different cohorts, consistently outperforming models based on raw abundances in cross-dataset evaluation. Our approach provides a valuable framework for leveraging self-supervised representation learning to overcome the data limitations inherent in microbiome research, potentially enabling more robust and generalizable machine learning applications in this field.


  • Open Access

    Diet shapes the gut microbiome: cross-sectional and longitudinal insights from the Human Phenotype Project

    by Tomer Segev, Daniel Barak, Liron Zahavi, Anastasia Godneva, Michal Rein, David Krongauz, Hagai Rossman, Adina Weinberger, Eran Segal

    Diet is a major environmental factor influencing the human gut microbiome. However, the effects of specific foods and dietary patterns on microbial composition, diversity, and function is not fully understood, limiting progress toward personalized dietary strategies. Leveraging 10,064 participants from the Human Phenotype Project with app-based diet logs and shotgun metagenomics, we predicted diet-microbiome associations at species-level resolution. Diet significantly predicted microbial diversity (Richness r=0.24, Shannon index r=0.22), the relative abundance of 664 of 724 species tested (91.5%, FDR<0.05), and 313 of 320 functional pathways (97.8%, FDR<0.05). Feature attribution identified distinct food-microbe links, including coffee with Lawsonibacter asaccharolyticus (r=0.41), yogurt with Streptococcus thermophilus (r=0.38), and milk with Bifidobacterium species (r=0.31-0.35). In parallel, broader dietary patterns, especially those related to the degree of food processing, emerged as predictors of microbial diversity and composition. Longitudinal analyses on test set data showed that models remained predictive over two and four years (R²=0.73-0.79), with significant associations between observed and predicted changes in 115 and 92 species, respectively. Finally, we developed an exploratory analysis for suggesting personalized dietary interventions with predicted microbiome shift effects that are associated with improvements in key clinical biomarkers such as triglycerides. Overall, our findings establish diet as a key modulator of microbiome composition, diversity, and function, and highlight its potential for guiding personalized interventions.


  • Top Cited Papers

    Deep phenotyping of health–disease continuum in the Human Phenotype Project

    by Lee Reicher, Smadar Shilo, Anastasia Godneva, Guy Lutsker, Liron Zahavi, Saar Shoer, David Krongauz, Michal Rein, Sarah Kohn, Tomer Segev, et al.

    The Human Phenotype Project (HPP) is a large-scale deep-phenotype prospective cohort. To date, approximately 28,000 participants have enrolled, with more than 13,000 completing their initial visit. The project is aimed at identifying novel molecular signatures with diagnostic, prognostic and therapeutic value, and at developing artificial intelligence (AI)-based predictive models for disease onset and progression. The HPP includes longitudinal profiling encompassing medical history, lifestyle and nutrition, anthropometrics, blood tests, continuous glucose and sleep monitoring, imaging and multi-omics data, including genetics, transcriptomics, microbiome (gut, vaginal and oral), metabolomics and immune profiling. Analysis of these data highlights the variation of phenotypes with age and ethnicity and unravels molecular signatures of disease by comparison with matched healthy controls. Leveraging extensive dietary and lifestyle data, we identify associations between lifestyle factors and health outcomes. Finally, we present a multi-modal foundation AI model, trained using self-supervised learning on diet and continuous-glucose-monitoring data, that outperforms existing methods in predicting disease onset. This framework can be extended to integrate other modalities and act as a personalized digital twin. In summary, we present a deeply phenotyped cohort that serves as a platform for advancing biomarker discovery, enabling the development of multi-modal AI models and personalized medicine approaches.


  • Open Access

    Accurate estimation of intraspecific microbial gene content variation in metagenomic data with MIDAS v3 and StrainPGC

    by Byron J. Smith, Chunyu Zhao, Veronika Dubinkina, Xiaofan Jin, Liron Zahavi, Saar Shoer, Jacqueline Moltzau-Anderson, Eran Segal, Katherine S. Pollard

    Metagenomics has greatly expanded our understanding of the human gut microbiome by revealing a vast diversity of bacterial species within and across individuals. Even within a single species, different strains can have highly divergent gene content, affecting traits such as antibiotic resistance, metabolism, and virulence. Methods that harness metagenomic data to resolve strain-level differences in functional potential are crucial for understanding the causes and consequences of this intraspecific diversity. The enormous size of pangenome references, strain mixing within samples, and inconsistent sequencing depth present challenges for existing tools that analyze samples one at a time. To address this gap, we updated the MIDAS pangenome profiler, now released as version 3, and developed StrainPGC, an approach to strain-specific gene content estimation that combines strain tracking and correlations across multiple samples. We validate our integrated analysis using a complex synthetic community of strains from the human gut and find that StrainPGC outperforms existing approaches. Analyzing a large, publicly available metagenome collection from inflammatory bowel disease patients and healthy controls, we catalog the functional repertoires of thousands of strains across hundreds of species, capturing extensive diversity missing from reference databases. Finally, we apply StrainPGC to metagenomes from a clinical trial of fecal microbiota transplantation for the treatment of ulcerative colitis. We identify twoEscherichia colistrains, from two different donors, that are both frequently transmitted to patients but have notable differences in functional potential. StrainPGC and MIDAS v3 together enable precise, intraspecific pangenomic investigations using large collections of metagenomic data without microbial isolation or de novo assembly.


  • Phenome-wide associations of human aging uncover sex-specific dynamics

    by Lee Reicher, Noam Bar, Anastasia Godneva, Yotam Reisner, Liron Zahavi, Nir Shahaf, Raja Dhir, Adina Weinberger, Eran Segal

    Aging varies significantly among individuals of the same chronological age, indicating that biological age (BA), estimated from molecular and physiological biomarkers, may better reflect aging. Prior research has often ignored sex-specific differences in aging patterns and mainly focused on aging biomarkers from a single data modality. Here we analyze a deeply phenotyped longitudinal cohort (10K project, Israel) of 10,000 healthy individuals aged 40–70 years that includes clinical, physiological, behavioral, environmental and multiomic parameters. Follow-up visits are scheduled every 2 years for a total of 25 years. We devised machine learning models of chronological age and computed biological aging scores that represented diverse physiological systems, revealing different aging patterns among sexes. Higher BA scores were associated with a higher prevalence of age-related medical conditions, highlighting the clinical relevance of these scores. Our analysis revealed system-specific aging dynamics and the potential of deeply phenotyped cohorts to accelerate improvements in our understanding of chronic diseases. Our findings present a more holistic view of the aging process, and lay the foundation for personalized medical prevention strategies.


  • Open Access

    Accurate estimation of intraspecific microbial gene content variation in metagenomic data with MIDAS v3 and StrainPGC

    by Byron J. Smith, Chunyu Zhao, Veronika Dubinkina, Xiaofan Jin, Liron Zahavi, Saar Shoer, Jacqueline Moltzau-Anderson, Eran Segal, Katherine S. Pollard

    Metagenomics has greatly expanded our understanding of the human gut microbiome by revealing a vast diversity of bacterial species within and across individuals. Even within a single species, different strains can have highly divergent gene content, affecting traits such as antibiotic resistance, metabolism, and virulence. Methods that harness metagenomic data to resolve strain-level differences in functional potential are crucial for understanding the causes and consequences of this intraspecific diversity. The enormous size of pangenome references, strain mixing within samples, and inconsistent sequencing depth present challenges for existing tools that analyze samples one at a time. To address this gap, we updated the MIDAS pangenome profiler, now released as version 3, and developed StrainPGC, an approach to strain-specific gene content estimation that combines strain tracking and correlations across multiple samples. We validate our integrated analysis using a complex synthetic community of strains from the human gut and find that StrainPGC outperforms existing approaches. Analyzing a large, publicly available metagenome collection from inflammatory bowel disease patients and healthy controls, we catalog the functional repertoires of thousands of strains across hundreds of species, capturing extensive diversity missing from reference databases. Finally, we apply StrainPGC to metagenomes from a clinical trial of fecal microbiota transplantation for the treatment of ulcerative colitis. We identify two Escherichia coli strains from two different donors that are both frequently transmitted to patients, but have notable differences in functional potential. StrainPGC and MIDAS v3 together enable precise, intraspecific pangenomic investigations using large collections of metagenomic data without microbial isolation or de novo assembly.


  • Bacterial SNPs in the human gut microbiome associate with host BMI

    by Liron Zahavi, Amit Lavon, Lee Reicher, Saar Shoer, Anastasia Godneva, Sigal Leviatan, Michal Rein, Omer Weissbrod, Adina Weinberger, Eran Segal

    Genome-wide association studies (GWASs) have provided numerous associations between human single-nucleotide polymorphisms (SNPs) and health traits. Likewise, metagenome-wide association studies (MWASs) between bacterial SNPs and human traits can suggest mechanistic links, but very few such studies have been done thus far. In this study, we devised an MWAS framework to detect SNPs and associate them with host phenotypes systematically. We recruited and obtained gut metagenomic samples from a cohort of 7,190 healthy individuals and discovered 1,358 statistically significant associations between a bacterial SNP and host body mass index (BMI), from which we distilled 40 independent associations. Most of these associations were unexplained by diet, medications or physical exercise, and 17 replicated in a geographically independent cohort. We uncovered BMI-associated SNPs in 27 bacterial species, and 12 of them showed no association by standard relative abundance analysis. We revealed a BMI association of an SNP in a potentially inflammatory pathway of Bilophila wadsworthia as well as of a group of SNPs in a region coding for energy metabolism functions in a Faecalibacterium prausnitzii genome. Our results demonstrate the importance of considering nucleotide-level diversity in microbiome studies and pave the way toward improved understanding of interpersonal microbiome differences and their potential health implications.