An atlas of protein homo-oligomerization across domains of life.

Protein structures are essential to understanding cellular processes in molecular detail. While advances in artificial intelligence revealed the tertiary structure of proteins at scale, their quaternary structure remains mostly unknown. We devise a scalable strategy based on AlphaFold2 to predict homo-oligomeric assemblies across four proteomes spanning the tree of life. Our results suggest that approximately 45% of an archaeal proteome and a bacterial proteome and 20% of two eukaryotic proteomes form homomers. Our predictions accurately capture protein homo-oligomerization, recapitulate megadalton complexes, and unveil hundreds of homo-oligomer types, including three confirmed experimentally by structure determination. Integrating these datasets with omics information suggests that a majority of known protein complexes are symmetric. Finally, these datasets provide a structural context for interpreting disease mutations and reveal coiled-coil regions as major enablers of quaternary structure evolution in human. Our strategy is applicable to any organism and provides a comprehensive view of homo-oligomerization in proteomes.

Opportunities and challenges in design and optimization of protein function.

The field of protein design has made remarkable progress over the past decade. Historically, the low reliability of purely structure-based design methods limited their application, but recent strategies that combine structure-based and sequence-based calculations, as well as machine learning tools, have dramatically improved protein engineering and design. In this Review, we discuss how these methods have enabled the design of increasingly complex structures and therapeutically relevant activities. Additionally, protein optimization methods have improved the stability and activity of complex eukaryotic proteins. Thanks to their increased reliability, computational design methods have been applied to improve therapeutics and enzymes for green chemistry and have generated vaccine antigens, antivirals and drug-delivery nano-vehicles. Moreover, the high success of design methods reflects an increased understanding of basic rules that govern the relationships among protein sequence, structure and function. However, de novo design is still limited mostly to α-helix bundles, restricting its potential to generate sophisticated enzymes and diverse protein and small-molecule binders. Designing complex protein structures is a challenging but necessary next step if we are to realize our objective of generating new-to-nature activities.

Computational design of soluble and functional membrane protein analogues.

De novo design of complex protein folds using solely computational means remains a substantial challenge1. Here we use a robust deep learning pipeline to design complex folds and soluble analogues of integral membrane proteins. Unique membrane topologies, such as those from G-protein-coupled receptors2, are not found in the soluble proteome, and we demonstrate that their structural features can be recapitulated in solution. Biophysical analyses demonstrate the high thermal stability of the designs, and experimental structures show remarkable design accuracy. The soluble analogues were functionalized with native structural motifs, as a proof of concept for bringing membrane protein functions to the soluble proteome, potentially enabling new approaches in drug discovery. In summary, we have designed complex protein topologies and enriched them with functionalities from membrane proteins, with high experimental success rates, leading to a de facto expansion of the functional soluble fold space.

Genomic analysis of diet composition finds novel loci and associations with health and lifestyle.

We conducted genome-wide association studies (GWAS) of relative intake from the macronutrients fat, protein, carbohydrates, and sugar in over 235,000 individuals of European ancestries. We identified 21 unique, approximately independent lead SNPs. Fourteen lead SNPs are uniquely associated with one macronutrient at genome-wide significance (P < 5 × 10-8), while five of the 21 lead SNPs reach suggestive significance (P < 1 × 10-5) for at least one other macronutrient. While the phenotypes are genetically correlated, each phenotype carries a partially unique genetic architecture. Relative protein intake exhibits the strongest relationships with poor health, including positive genetic associations with obesity, type 2 diabetes, and heart disease (rg ≈ 0.15-0.5). In contrast, relative carbohydrate and sugar intake have negative genetic correlations with waist circumference, waist-hip ratio, and neighborhood deprivation (|rg| ≈ 0.1-0.3) and positive genetic correlations with physical activity (rg ≈ 0.1 and 0.2). Relative fat intake has no consistent pattern of genetic correlations with poor health but has a negative genetic correlation with educational attainment (rg ≈-0.1). Although our analyses do not allow us to draw causal conclusions, we find no evidence of negative health consequences associated with relative carbohydrate, sugar, or fat intake. However, our results are consistent with the hypothesis that relative protein intake plays a role in the etiology of metabolic dysfunction.

On the evolution and physiology of cable bacteria.

Cable bacteria of the family Desulfobulbaceae form centimeter-long filaments comprising thousands of cells. They occur worldwide in the surface of aquatic sediments, where they connect sulfide oxidation with oxygen or nitrate reduction via long-distance electron transport. In the absence of pure cultures, we used single-filament genomics and metagenomics to retrieve draft genomes of 3 marine Candidatus Electrothrix and 1 freshwater Ca. Electronema species. These genomes contain >50% unknown genes but still share their core genomic makeup with sulfate-reducing and sulfur-disproportionating Desulfobulbaceae, with few core genes lost and 212 unique genes (from 197 gene families) conserved among cable bacteria. Last common ancestor analysis indicates gene divergence and lateral gene transfer as equally important origins of these unique genes. With support from metaproteomics of a Ca. Electronema enrichment, the genomes suggest that cable bacteria oxidize sulfide by reversing the canonical sulfate reduction pathway and fix CO2 using the Wood-Ljungdahl pathway. Cable bacteria show limited organotrophic potential, may assimilate smaller organic acids and alcohols, fix N2, and synthesize polyphosphates and polyglucose as storage compounds; several of these traits were confirmed by cell-level experimental analyses. We propose a model for electron flow from sulfide to oxygen that involves periplasmic cytochromes, yet-unidentified conductive periplasmic fibers, and periplasmic oxygen reduction. This model proposes that an active cable bacterium gains energy in the anodic, sulfide-oxidizing cells, whereas cells in the oxic zone flare off electrons through intense cathodic oxygen respiration without energy conservation; this peculiar form of multicellularity seems unparalleled in the microbial world.

Genetic instrumental variable regression: Explaining socioeconomic and health outcomes in nonexperimental data.

Identifying causal effects in nonexperimental data is an enduring challenge. One proposed solution that recently gained popularity is the idea to use genes as instrumental variables [i.e., Mendelian randomization (MR)]. However, this approach is problematic because many variables of interest are genetically correlated, which implies the possibility that many genes could affect both the exposure and the outcome directly or via unobserved confounding factors. Thus, pleiotropic effects of genes are themselves a source of bias in nonexperimental data that would also undermine the ability of MR to correct for endogeneity bias from nongenetic sources. Here, we propose an alternative approach, genetic instrumental variable (GIV) regression, that provides estimates for the effect of an exposure on an outcome in the presence of pleiotropy. As a valuable byproduct, GIV regression also provides accurate estimates of the chip heritability of the outcome variable. GIV regression uses polygenic scores (PGSs) for the outcome of interest which can be constructed from genome-wide association study (GWAS) results. By splitting the GWAS sample for the outcome into nonoverlapping subsamples, we obtain multiple indicators of the outcome PGSs that can be used as instruments for each other and, in combination with other methods such as sibling fixed effects, can address endogeneity bias from both pleiotropy and the environment. In two empirical applications, we demonstrate that our approach produces reasonable estimates of the chip heritability of educational attainment (EA) and show that standard regression and MR provide upwardly biased estimates of the effect of body height on EA.

Ventral Capsule/Ventral Striatum Stimulation in Obsessive-Compulsive Disorder: Toward a Unified Connectomic Target for Deep Brain Stimulation?

INTRODUCTION: Obsessive-compulsive disorder (OCD) is among the most disabling chronic psychiatric disorders and has a significant negative impact on multiple domains of quality of life. Deep brain stimulation (DBS) is a treatment option for severe therapy-resistant OCD. OBJECTIVE: To provide a detailed clinical description and treatment outcome analysis in a cohort of eight refractory OCD patients receiving ventral capsule/ventral striatum (VC/VS) stimulation with the intention to validate discriminating fiber bundles previously associated with clinical response. MATERIALS AND METHODS: The primary outcome measure (the Yale-Brown Obsessive Compulsive Scale [Y-BOCS]) and secondary outcomes depressive symptoms, anxiety, and quality of life were retrospectively analyzed. DBS leads were warped into standard stereotactic space. A normative connectome was used to identify the neural network associated with clinical outcome. RESULTS: With a median stimulation duration of 26 months, patients exhibited a mean Y-BOCS reduction of 10.5 resulting in a response rate of 63%. Modulation of a fiber bundle traversing the anterior limb of the internal capsule (ALIC) was associated with Y-BOCS reduction. This fiber bundle connected the frontal regions to the subthalamic nucleus (STN) and was functionally identified as the hyperdirect pathway of the basal ganglia circuitry. CONCLUSION: Our findings show that in VC/VS stimulation, the neural network associated with clinical outcome shows overlap with that of previously described for other targets namely the anterior limb of the internal capsula, the nucleus accumbens, or the STN, which supports the evolvement from the concept of an optimal gray matter target to conceiving the target as part of a symptom modulating network.

Long-term treatment with metformin in type 2 diabetes and vitamin D levels: A post-hoc analysis of a randomized placebo-controlled trial.

AIMS: To study the effects of metformin, as compared to placebo, on serum levels of vitamin D (25-hydroxyvitamin D [25(OH)D]) in patients with advanced type 2 diabetes. MATERIALS AND METHODS: In the HOME trial, a randomized placebo-controlled trial, 390 insulin-treated patients with type 2 diabetes were treated with 850 mg metformin or placebo thrice daily for 52 months. In a post-hoc analysis, we examined changes in the combined levels of 25(OH)D2 and 25(OH)D3 at 4 and 16 months during the study. RESULTS: Mean combined 25(OH)D at baseline was 68.2 nmoL/L (95% confidence interval [CI]: 65.5-71.1). In mixed model analysis, metformin, as compared to placebo, had no effect on 25(OH)D levels during 16 months (coefficient: 1.002 per month, multiplicative model; 95% CI: 0.998-1.006, P = .30). Metformin was associated with a small increase of 25(OH)D2 (coefficient: 1.012 per month; 95% CI: 1.003-1.021, P = .008). However, 25(OH)D2 is only a very small fraction (3%) of 25(OH)D. Seasonal variation had the biggest impact on 25(OH)D levels. Vitamin B12 levels were not associated with the levels of 25(OH)D. CONCLUSION: Metformin had no effect on serum 25(OH)D during 16 months in the setting of a clinical randomized controlled trial in patients with type 2 diabetes. Our results show that metformin doesn't lead to vitamin D deficiency.

No evidence for cerebellar abnormality in adults with developmental dyslexia.

Developmental dyslexia is commonly believed to result from a deficiency in the recognition and processing of speech sounds. According to the cerebellar deficit hypothesis, this phonological deficit is caused by deficient cerebellar function. In the current study, 26 adults with developmental dyslexia and 25 non-dyslexic participants underwent testing of reading-related skills, cerebellar functions, and MRI scanning of the brain. Anatomical assessment of the cerebellum was conducted with voxel-based morphometry. Behavioural evidence, that was indicative of impaired cerebellar function, was found to co-occur with reading impairments in the dyslexic subjects, but a causal relation between the two was not observed. No differences in local grey matter volume, nor in structure-function relationships within the cerebellum were found between the two groups. Possibly, the observed behavioural pattern is due to aberrant white matter connectivity. In conclusion, no support for the cerebellar deficit hypothesis or the presence of anatomical differences of the cerebellum in adults with developmental dyslexia was found.

Safety, efficacy and repeatability of a novel house dust mite allergen challenge technique in the Fraunhofer allergen challenge chamber.

BACKGROUND: Efficacy testing of immunotherapy in field studies is often hampered by variation of airborne allergens. Standardized allergen exposure in challenge chamber settings might be an alternative. Therefore, we developed a universal technique to create an atmosphere loaded with allergen particles of adjustable size from aqueous solutions of licensed allergen extracts. OBJECTIVE: The aim of this study was to apply this technique and test the safety and efficacy of challenges with house dust mite (HDM) allergen in the Fraunhofer allergen challenge chamber. METHODS: Aerosol particles carrying HDM allergen were produced by spray-drying of an aqueous solution containing HDM allergen and lactose. In a monocenter, placebo-controlled, single-blind, dose-escalation pilot study, 18 subjects with perennial allergic rhinitis and sensitization to HDM were exposed to HDM allergen for 4 h at either 250, 500, 1000 SQE/m3 or lactose alone (0 SQE/m3 ) 7 days apart. The dose of 500 SQE/m³ was repeated to investigate reproducibility. Total nasal symptom score (TNSS) was the primary endpoint. RESULTS: Exposure to HDM increased TNSS (mean ± SD) to 3.4 ± 1.8, 3.3 ± 2.1, and 3.6 ± 2.0 at 250, 500 and 1000 SQE/m3 , respectively, while lactose alone did not change TNSS (0.7 ± 0.6). The results were reproducible at 500 SQE/m3 . Pulmonary function and adverse event frequency did not change with escalation of allergen dose. CONCLUSION: This HDM allergen particle generation is safe, specific and reproducible and can therefore be used for efficacy testing of immunotherapy and for basic clinical research.

Computational design of soluble functional analogues of integral membrane proteins.

De novo design of complex protein folds using solely computational means remains a significant challenge. Here, we use a robust deep learning pipeline to design complex folds and soluble analogues of integral membrane proteins. Unique membrane topologies, such as those from GPCRs, are not found in the soluble proteome and we demonstrate that their structural features can be recapitulated in solution. Biophysical analyses reveal high thermal stability of the designs and experimental structures show remarkable design accuracy. The soluble analogues were functionalized with native structural motifs, standing as a proof-of-concept for bringing membrane protein functions to the soluble proteome, potentially enabling new approaches in drug discovery. In summary, we designed complex protein topologies and enriched them with functionalities from membrane proteins, with high experimental success rates, leading to a de facto expansion of the functional soluble fold space.

De novo protein design by inversion of the AlphaFold structure prediction network.

De novo protein design enhances our understanding of the principles that govern protein folding and interactions, and has the potential to revolutionize biotechnology through the engineering of novel protein functionalities. Despite recent progress in computational design strategies, de novo design of protein structures remains challenging, given the vast size of the sequence-structure space. AlphaFold2 (AF2), a state-of-the-art neural network architecture, achieved remarkable accuracy in predicting protein structures from amino acid sequences. This raises the question whether AF2 has learned the principles of protein folding sufficiently for de novo design. Here, we sought to answer this question by inverting the AF2 network, using the prediction weight set and a loss function to bias the generated sequences to adopt a target fold. Initial design trials resulted in de novo designs with an overrepresentation of hydrophobic residues on the protein surface compared to their natural protein family, requiring additional surface optimization. In silico validation of the designs showed protein structures with the correct fold, a hydrophilic surface and a densely packed hydrophobic core. In vitro validation showed that 7 out of 39 designs were folded and stable in solution with high melting temperatures. In summary, our design workflow solely based on AF2 does not seem to fully capture basic principles of de novo protein design, as observed in the protein surface's hydrophobic vs. hydrophilic patterning. However, with minimal post-design intervention, these pipelines generated viable sequences as assessed experimental characterization. Thus, such pipelines show the potential to contribute to solving outstanding challenges in de novo protein design.

Indications for a genetic basis for big bacteria and description of the giant cable bacterium Candidatus Electrothrix gigas sp. nov.

Bacterial cells can vary greatly in size, from a few hundred nanometers to hundreds of micrometers in diameter. Filamentous cable bacteria also display substantial size differences, with filament diameters ranging from 0.4 to 8 µm. We analyzed the genomes of cable bacterium filaments from 11 coastal environments of which the resulting 23 new genomes represent 10 novel species-level clades of Candidatus Electrothrix and two clades that putatively represent novel genus-level diversity. Fluorescence in situ hybridization with a species-level probe showed that large-sized cable bacteria belong to a novel species with the proposed name Ca. Electrothrix gigas. Comparative genome analysis suggests genes that play a role in the construction or functioning of large cable bacteria cells: the genomes of Ca. Electrothrix gigas encode a novel actin-like protein as well as a species-specific gene cluster encoding four putative pilin proteins and a putative type II secretion platform protein, which are not present in other cable bacteria. The novel actin-like protein was also found in a number of other giant bacteria, suggesting there could be a genetic basis for large cell size. This actin-like protein (denoted big bacteria protein, Bbp) may have a function analogous to other actin proteins in cell structure or intracellular transport. We contend that Bbp may help overcome the challenges of diffusion limitation and/or morphological complexity presented by the large cells of Ca. Electrothrix gigas and other giant bacteria. IMPORTANCE In this study, we substantially expand the known diversity of marine cable bacteria and describe cable bacteria with a large diameter as a novel species with the proposed name Candidatus Electrothrix gigas. In the genomes of this species, we identified a gene that encodes a novel actin-like protein [denoted big bacteria protein (Bbp)]. The bbp gene was also found in a number of other giant bacteria, predominantly affiliated to Desulfobacterota and Gammaproteobacteria, indicating that there may be a genetic basis for large cell size. Thus far, mostly structural adaptations of giant bacteria, vacuoles, and other inclusions or organelles have been observed, which are employed to overcome nutrient diffusion limitation in their environment. In analogy to other actin proteins, Bbp could fulfill a structural role in the cell or potentially facilitate intracellular transport.

Cable bacteria with electric connection to oxygen attract flocks of diverse bacteria.

Cable bacteria are centimeter-long filamentous bacteria that conduct electrons via internal wires, thus coupling sulfide oxidation in deeper, anoxic sediment with oxygen reduction in surface sediment. This activity induces geochemical changes in the sediment, and other bacterial groups appear to benefit from the electrical connection to oxygen. Here, we report that diverse bacteria swim in a tight flock around the anoxic part of oxygen-respiring cable bacteria and disperse immediately when the connection to oxygen is disrupted (by cutting the cable bacteria with a laser). Raman microscopy shows that flocking bacteria are more oxidized when closer to the cable bacteria, but physical contact seems to be rare and brief, which suggests potential transfer of electrons via unidentified soluble intermediates. Metagenomic analysis indicates that most of the flocking bacteria appear to be aerobes, including organotrophs, sulfide oxidizers, and possibly iron oxidizers, which might transfer electrons to cable bacteria for respiration. The association and close interaction with such diverse partners might explain how oxygen via cable bacteria can affect microbial communities and processes far into anoxic environments.

Prognostic Value of Histopathological Thrombus Age in Large Vessel Occlusion-Related Stroke.

INTRODUCTION: Acute mechanical thrombectomy (MT) is the preferred treatment for large vessel occlusion-related stroke. Histopathological research on the obtained occlusive embolic thrombus may provide information regarding the aetiology and pathology of the lesion to predict prognosis and propose possible future acute ischaemic stroke therapy. METHODS: A total of 75 consecutive patients who presented to the Amphia Hospital with acute large vessel occlusion-related stroke and underwent MT were included in the study. The obtained thrombus materials were subjected to standard histopathological examination. Based on histological criteria, they were considered fresh (<1 day old) or old (>1 day old). Patients were followed for 2 years for documentation of all-cause mortality. RESULTS: Thrombi were classified as fresh in 40 patients (53%) and as older in 35 patients (47%). Univariate Cox regression analysis showed that thrombus age, National Institutes of Health Stroke Scale at hospital admission, and patient age were associated with long-term mortality (p < 0.1). Multivariable Cox hazards and Kaplan-Meier analysis demonstrated that after extensive adjustment for clinical and procedural variables, thrombus age persisted in being independently associated with higher long-term mortality (hazard ratio: 3.34; p = 0.038, log-rank p = 0.013). CONCLUSION: In this study, older thromboemboli are responsible for almost half of acute large ischaemic strokes. Moreover, the presence of an old thrombus is an independent predictor of mortality in acute large vessel occlusion-related stroke. More research is warranted regarding future therapies based on thrombus composition.

Revealing the queer-spectrum in STEM through robust demographic data collection in undergraduate engineering and computer science courses at four institutions.

Queer identities are often ignored in diversity initiatives, yet there is a growing body of research that describes notable heterosexist and gender-normative expectations in STEM that lead to unsupportive and discriminatory environments and to the lower persistence of queer individuals. Research on the experiences of queer-spectrum individuals is limited by current demographic practices. In surveys that are queer-inclusive there is no consensus on best practices, and individuals with queer genders and queer sexual, romantic, and related orientations are often lumped together in a general category (e.g. LGBTQ+). We developed two queer-inclusive demographics questions and administered them as part of a larger study in undergraduate engineering and computer science classes (n = 3698), to determine which of three survey types for gender (conventional, queered, open-ended) provided the most robust data and compared responses to national data to determine if students with queer genders and/or queer sexual, romantic, and related orientations were underrepresented in engineering and computer science programs. The gender survey with queer-identity options provided the most robust data, as measured by higher response rates and relatively high rates of disclosing queer identities. The conventional survey (male, female, other) had significantly fewer students disclose queer identities, and the open-ended survey had a significantly higher non-response rate. Allowing for multiple responses on the survey was important: 78% of those with queer gender identities and 9% of those with queer sexual, romantic and related orientations selected multiple identities within the same survey question. Queer students in our study were underrepresented relative to national data. Students who disclosed queer gender identities were 7/100ths of the expected number, and those with queer orientations were under-represented by one-quarter. Further work developing a research-based queered demographics instrument is needed for larger-scale changes in demographics practices, which will help others identify and address barriers that queer-spectrum individuals face in STEM.

"It's completely erasure": A Qualitative Exploration of Experiences of Transgender, Nonbinary, Gender Nonconforming, and Questioning Students in Biology Courses.

Biology is the study of the diversity of life, which includes diversity in sex, gender, and sexual, romantic, and related orientations. However, a small body of literature suggests that undergraduate biology courses focus on only a narrow representation of this diversity (binary sexes, heterosexual orientations, etc.). In this study, we interviewed students with queer genders to understand the messages about sex, gender, and orientation they encountered in biology and the impact of these messages on them. We found five overarching themes in these interviews. Students described two narratives about sex, gender, and orientation in their biology classes that made biology implicitly exclusionary. These narratives harmed students by impacting their sense of belonging, career preparation, and interest in biology content. However, students employed a range of resilience strategies to resist these harms. Finally, students described the currently unrealized potential for biology and biology courses to validate queer identities by representing the diversity in sex and orientation in biology. We provide teaching suggestions derived from student interviews for making biology more queer-inclusive.

Resource profile and user guide of the Polygenic Index Repository.

Polygenic indexes (PGIs) are DNA-based predictors. Their value for research in many scientific disciplines is growing rapidly. As a resource for researchers, we used a consistent methodology to construct PGIs for 47 phenotypes in 11 datasets. To maximize the PGIs' prediction accuracies, we constructed them using genome-wide association studies-some not previously published-from multiple data sources, including 23andMe and UK Biobank. We present a theoretical framework to help interpret analyses involving PGIs. A key insight is that a PGI can be understood as an unbiased but noisy measure of a latent variable we call the 'additive SNP factor'. Regressions in which the true regressor is this factor but the PGI is used as its proxy therefore suffer from errors-in-variables bias. We derive an estimator that corrects for the bias, illustrate the correction, and make a Python tool for implementing it publicly available.

Contribution of the left and right inferior frontal gyrus in recovery from aphasia. A functional MRI study in stroke patients with preserved hemodynamic responsiveness.

The relative contribution of dominant and non-dominant language networks to recovery from aphasia is a matter of debate. We assessed with functional magnetic resonance imaging (fMRI) to what extent the left and right hemispheres are associated with recovery from aphasia after stroke. fMRI with three language tasks was performed in 13 aphasic stroke patients and in 13 healthy subjects. Severity of aphasia was examined within 2 months after stroke and after at least 1 year. Recovery of naming ability and scores on the Token Test were correlated with data from fMRI in the chronic phase. A breath-hold paradigm was used to investigate hemodynamic responsiveness. Overall language performance in the chronic phase correlated with higher relative activation of left compared to right perisylvian areas. Recovery of naming ability was positively correlated with activation in the left inferior frontal gyrus (IFG) for semantic decision and verb generation. Recovery on the Token Test was positively correlated with activation in both left and right IFG during semantic decision and verb generation. Hemodynamic response to the breath-hold task was similar in patients and controls. Our study suggests that in the chronic stage after stroke left IFG activity is associated with improvement of picture naming and sentence comprehension, whereas activity in the right IFG may reflect up-regulation of non-linguistic cognitive processing. Altered hemodynamic responsiveness seems an unlikely confounder in the interpretations of fMRI results.

Long-term treatment with metformin in type 2 diabetes and methylmalonic acid: Post hoc analysis of a randomized controlled 4.3year trial.

AIMS: Metformin treatment is associated with a decrease of serum vitamin B12, but whether this reflects tissue B12 deficiency is controversial. We studied the effects of metformin on serum levels of methylmalonic acid (MMA), a biomarker for tissue B12 deficiency, and on onset or progression of neuropathy. METHODS: In the HOME trial, 390 insulin-treated patients with type 2 diabetes were treated with metformin or placebo for 52months. In a post hoc analysis, we analyzed the association between metformin, MMA and a validated Neuropathy Score (NPS). RESULTS: Metformin vs placebo increased MMA at the end of the study (95%CI: 0.019 to 0.055, p=0.001). Mediation analysis showed that the effect of metformin on the NPS consisted of a beneficial effect through lowering HbA1c (-0.020 per gram year) and an adverse effect through increasing MMA (0.042 per gram year), resulting in a non-significant net effect (0.032 per gram year, 95% CI: -0.121 to 0.182, p=0.34). CONCLUSION: Metformin not only reduces serum levels of B12, but also progressively increases serum MMA. The increase of MMA in metformin users was associated with significant worsening of the NPS. These results provide further support that metformin-related B12 deficiency is clinically relevant. Monitoring of B12 in users of metformin should be considered.