DeepLoc 2.1: multi-label membrane protein type prediction using protein language models.

DeepLoc 2.0 is a popular web server for the prediction of protein subcellular localization and sorting signals. Here, we introduce DeepLoc 2.1, which additionally classifies the input proteins into the membrane protein types Transmembrane, Peripheral, Lipid-anchored and Soluble. Leveraging pre-trained transformer-based protein language models, the server utilizes a three-stage architecture for sequence-based, multi-label predictions. Comparative evaluations with other established tools on a test set of 4933 eukaryotic protein sequences, constructed following stringent homology partitioning, demonstrate state-of-the-art performance. Notably, DeepLoc 2.1 outperforms existing models, with the larger ProtT5 model exhibiting a marginal advantage over the ESM-1B model. The web server is available at https://services.healthtech.dtu.dk/services/DeepLoc-2.1.

DeepPeptide predicts cleaved peptides in proteins using conditional random fields.

MOTIVATION: Peptides are ubiquitous throughout life and involved in a wide range of biological processes, ranging from neural signaling in higher organisms to antimicrobial peptides in bacteria. Many peptides are generated post-translationally by cleavage of precursor proteins and can thus not be detected directly from genomics data, as the specificities of the responsible proteases are often not completely understood. RESULTS: We present DeepPeptide, a deep learning model that predicts cleaved peptides directly from the amino acid sequence. DeepPeptide shows both improved precision and recall for peptide detection compared to previous methodology. We show that the model is capable of identifying peptides in underannotated proteomes. AVAILABILITY AND IMPLEMENTATION: DeepPeptide is available online at ku.biolib.com/DeepPeptide.

Deorphanizing Peptides Using Structure Prediction.

Many endogenous peptides rely on signaling pathways to exert their function, but identifying their cognate receptors remains a challenging problem. We investigate the use of AlphaFold-Multimer complex structure prediction together with transmembrane topology prediction for peptide deorphanization. We find that AlphaFold's confidence metrics have strong performance for prioritizing true peptide-receptor interactions. In a library of 1112 human receptors, the method ranks true receptors in the top percentile on average for 11 benchmark peptide-receptor pairs.

Body-mass index and diabetes risk in 57 low-income and middle-income countries: a cross-sectional study of nationally representative, individual-level data in 685 616 adults.

BACKGROUND: The prevalence of overweight, obesity, and diabetes is rising rapidly in low-income and middle-income countries (LMICs), but there are scant empirical data on the association between body-mass index (BMI) and diabetes in these settings. METHODS: In this cross-sectional study, we pooled individual-level data from nationally representative surveys across 57 LMICs. We identified all countries in which a WHO Stepwise Approach to Surveillance (STEPS) survey had been done during a year in which the country fell into an eligible World Bank income group category. For LMICs that did not have a STEPS survey, did not have valid contact information, or declined our request for data, we did a systematic search for survey datasets. Eligible surveys were done during or after 2008; had individual-level data; were done in a low-income, lower-middle-income, or upper-middle-income country; were nationally representative; had a response rate of 50% or higher; contained a diabetes biomarker (either a blood glucose measurement or glycated haemoglobin [HbA1c]); and contained data on height and weight. Diabetes was defined biologically as a fasting plasma glucose concentration of 7·0 mmol/L (126·0 mg/dL) or higher; a random plasma glucose concentration of 11·1 mmol/L (200·0 mg/dL) or higher; or a HbA1c of 6·5% (48·0 mmol/mol) or higher, or by self-reported use of diabetes medication. We included individuals aged 25 years or older with complete data on diabetes status, BMI (defined as normal [18·5-22·9 kg/m2], upper-normal [23·0-24·9 kg/m2], overweight [25·0-29·9 kg/m2], or obese [≥30·0 kg/m2]), sex, and age. Countries were categorised into six geographical regions: Latin America and the Caribbean, Europe and central Asia, east, south, and southeast Asia, sub-Saharan Africa, Middle East and north Africa, and Oceania. We estimated the association between BMI and diabetes risk by multivariable Poisson regression and receiver operating curve analyses, stratified by sex and geographical region. FINDINGS: Our pooled dataset from 58 nationally representative surveys in 57 LMICs included 685 616 individuals. The overall prevalence of overweight was 27·2% (95% CI 26·6-27·8), of obesity was 21·0% (19·6-22·5), and of diabetes was 9·3% (8·4-10·2). In the pooled analysis, a higher risk of diabetes was observed at a BMI of 23 kg/m2 or higher, with a 43% greater risk of diabetes for men and a 41% greater risk for women compared with a BMI of 18·5-22·9 kg/m2. Diabetes risk also increased steeply in individuals aged 35-44 years and in men aged 25-34 years in sub-Saharan Africa. In the stratified analyses, there was considerable regional variability in this association. Optimal BMI thresholds for diabetes screening ranged from 23·8 kg/m2 among men in east, south, and southeast Asia to 28·3 kg/m2 among women in the Middle East and north Africa and in Latin America and the Caribbean. INTERPRETATION: The association between BMI and diabetes risk in LMICs is subject to substantial regional variability. Diabetes risk is greater at lower BMI thresholds and at younger ages than reflected in currently used BMI cutoffs for assessing diabetes risk. These findings offer an important insight to inform context-specific diabetes screening guidelines. FUNDING: Harvard T H Chan School of Public Health McLennan Fund: Dean's Challenge Grant Program.

SignalP: The Evolution of a Web Server.

SignalP ( https://services.healthtech.dtu.dk/services/SignalP-6.0/ ) is a very popular prediction method for signal peptides, the intrinsic signals that make proteins secretory. The SignalP web server has existed since 1995 and is now in its sixth major version. In this historical account, we (three authors who have taken part in the entire journey plus the first author of the latest version) describe the differences between the versions and discuss the various decisions taken along the way.

The Association Between Depressive Symptoms, Access to Diabetes Care, and Glycemic Control in Five Middle-Income Countries.

OBJECTIVE: The relationship between depression, diabetes, and access to diabetes care is established in high-income countries (HICs) but not in middle-income countries (MICs), where contexts and health systems differ and may impact this relationship. In this study, we investigate access to diabetes care for individuals with and without depressive symptoms in MICs. RESEARCH DESIGN AND METHODS: We analyzed pooled data from nationally representative household surveys across Brazil, Chile, China, Indonesia, and Mexico. Validated survey tools Center for Epidemiologic Studies Depression Scale Revised, Composite International Diagnostic Interview, Short Form, and Patient Health Questionnaire identified participants with depressive symptoms. Diabetes, defined per World Health Organization Package of Essential Noncommunicable Disease Interventions guidelines, included self-reported medication use and biochemical data. The primary focus was on tracking diabetes care progression through the stages of diagnosis, treatment, and glycemic control. Descriptive and multivariable logistic regression analyses, accounting for gender, age, education, and BMI, examined diabetes prevalence and care continuum progression. RESULTS: The pooled sample included 18,301 individuals aged 50 years and above; 3,309 (18.1%) had diabetes, and 3,934 (21.5%) exhibited depressive symptoms. Diabetes prevalence was insignificantly higher among those with depressive symptoms (28.9%) compared with those without (23.8%, P = 0.071). Co-occurrence of diabetes and depression was associated with increased odds of diabetes detection (odds ratio [OR] 1.398, P < 0.001) and treatment (OR 1.344, P < 0.001), but not with higher odds of glycemic control (OR 0.913, P = 0.377). CONCLUSIONS: In MICs, individuals aged 50 years and older with diabetes and depression showed heightened diabetes identification and treatment probabilities, unlike patterns seen in HICs. This underscores the unique interplay of these conditions in different income settings.

GraphPart: homology partitioning for biological sequence analysis.

When splitting biological sequence data for the development and testing of predictive models, it is necessary to avoid too-closely related pairs of sequences ending up in different partitions. If this is ignored, performance of prediction methods will tend to be overestimated. Several algorithms have been proposed for homology reduction, where sequences are removed until no too-closely related pairs remain. We present GraphPart, an algorithm for homology partitioning that divides the data such that closely related sequences always end up in the same partition, while keeping as many sequences as possible in the dataset. Evaluation of GraphPart on Protein, DNA and RNA datasets shows that it is capable of retaining a larger number of sequences per dataset, while providing homology separation on a par with reduction approaches.

The dynamic landscape of peptide activity prediction.

Peptides are known to possess a plethora of beneficial properties and activities: antimicrobial, anticancer, anti-inflammatory or the ability to cross the blood-brain barrier are only a few examples of their functional diversity. For this reason, bioinformaticians are constantly developing and upgrading models to predict their activity in silico, generating a steadily increasing number of available tools. Although these efforts have provided fruitful outcomes in the field, the vast and diverse amount of resources for peptide prediction can turn a simple prediction into an overwhelming searching process to find the optimal tool. This minireview aims at providing a systematic and accessible analysis of the complex ecosystem of peptide activity prediction, showcasing the variability of existing models for peptide assessment, their domain specialization and popularity. Moreover, we also assess the reproducibility of such bioinformatics tools and describe tendencies observed in their development. The list of tools is available under https://biogenies.info/peptide-prediction-list/.

SignalP 6.0 predicts all five types of signal peptides using protein language models.

Signal peptides (SPs) are short amino acid sequences that control protein secretion and translocation in all living organisms. SPs can be predicted from sequence data, but existing algorithms are unable to detect all known types of SPs. We introduce SignalP 6.0, a machine learning model that detects all five SP types and is applicable to metagenomic data.

Combining mass spectrometry and machine learning to discover bioactive peptides.

Peptides play important roles in regulating biological processes and form the basis of a multiplicity of therapeutic drugs. To date, only about 300 peptides in human have confirmed bioactivity, although tens of thousands have been reported in the literature. The majority of these are inactive degradation products of endogenous proteins and peptides, presenting a needle-in-a-haystack problem of identifying the most promising candidate peptides from large-scale peptidomics experiments to test for bioactivity. To address this challenge, we conducted a comprehensive analysis of the mammalian peptidome across seven tissues in four different mouse strains and used the data to train a machine learning model that predicts hundreds of peptide candidates based on patterns in the mass spectrometry data. We provide in silico validation examples and experimental confirmation of bioactivity for two peptides, demonstrating the utility of this resource for discovering lead peptides for further characterization and therapeutic development.

The effect of bearing and rearing a child on blood pressure: a nationally representative instrumental variable analysis of 444 611 mothers in India.

BACKGROUND: At the individual level, it is well known that pregnancies have a short-term effect on a woman's cardiovascular system and blood pressure. The long-term effect of having children on maternal blood pressure, however, is unknown. We thus estimated the causal effect of having children on blood pressure among mothers in India, a country with a history of high fertility rates. METHODS: We used nationally representative cross-sectional data from the 2015-16 India National Family and Health Survey (NFHS-4). The study population comprised 444 611 mothers aged 15-49 years. We used the sex of the first-born child as an instrumental variable (IV) for the total number of a woman's children. We estimated the effect of an additional child on systolic and diastolic blood pressure in IV (two-stage least squares) regressions. In additional analyses, we stratified the IV regressions by time since a mother last gave birth. Furthermore, we repeated our analyses using mothers' husbands and partners as the regression sample. RESULTS: On average, mothers had 2.7 children [standard deviation (SD): 1.5], a systolic blood pressure of 116.4 mmHg (SD: 14.4) and diastolic blood pressure of 78.5 mmHg (SD: 9.4). One in seven mothers was hypertensive. In conventional ordinary least squares regression, each child was associated with 0.42 mmHg lower systolic [95% confidence interval (CI): -0.46 to -0.39, P < 0.001] and 0.13 mmHg lower diastolic (95% CI: -0.15 to -0.11, P < 0.001) blood pressure. In the IV regressions, each child decreased a mother's systolic blood pressure by an average of 1.00 mmHg (95% CI: -1.26 to -0.74, P < 0.001) and diastolic blood pressure by an average of 0.35 mmHg (95% CI: -0.52 to -0.17, P < 0.001). These decreases were sustained over more than a decade after childbirth, with effect sizes slightly declining as the time since last birth increased. Having children did not influence blood pressure in men. CONCLUSIONS: Bearing and rearing a child decreases blood pressure among mothers in India.

Analysis of Attained Height and Diabetes Among 554,122 Adults Across 25 Low- and Middle-Income Countries.

OBJECTIVE: The prevalence of type 2 diabetes is rising rapidly in low-income and middle-income countries (LMICs), but the factors driving this rapid increase are not well understood. Adult height, in particular shorter height, has been suggested to contribute to the pathophysiology and epidemiology of diabetes and may inform how adverse environmental conditions in early life affect diabetes risk. We therefore systematically analyzed the association of adult height and diabetes across LMICs, where such conditions are prominent. RESEARCH DESIGN AND METHODS: We pooled individual-level data from nationally representative surveys in LMICs that included anthropometric measurements and diabetes biomarkers. We calculated odds ratios (ORs) for the relationship between attained adult height and diabetes using multilevel mixed-effects logistic regression models. We estimated ORs for the pooled sample, major world regions, and individual countries, in addition to stratifying all analyses by sex. We examined heterogeneity by individual-level characteristics. RESULTS: Our sample included 554,122 individuals across 25 population-based surveys. Average height was 161.7 cm (95% CI 161.2-162.3), and the crude prevalence of diabetes was 7.5% (95% CI 6.9-8.2). We found no relationship between adult height and diabetes across LMICs globally or in most world regions. When stratifying our sample by country and sex, we found an inverse association between adult height and diabetes in 5% of analyses (2 out of 50). Results were robust to alternative model specifications. CONCLUSIONS: Adult height is not associated with diabetes across LMICs. Environmental factors in early life reflected in attained adult height likely differ from those predisposing individuals for diabetes.

GLP-1 Receptor Agonists and the Path to Sustainable Obesity Care.

This Viewpoint discusses the optimal treatment duration of glucagon-like peptide-1 receptor agonists in people with obesity and the benefits of off-ramping, the tapering of these antiobesity medications following an initial treatment period.

Shorter Height is Associated with Diabetes in Women but not in Men: Nationally Representative Evidence from Namibia.

OBJECTIVE: This study aimed to test the hypothesis that attained adult height, as an indicator of childhood nutrition, is associated with diabetes in adulthood in Namibia, a country where stunting is highly prevalent. METHODS: Data from 1,898 women and 1,343 men aged 35 to 64 years included in the Namibia Demographic and Health Survey in 2013 were analyzed. Multiple logistic regression models were used to calculate odds ratios (ORs) and 95% CIs of having diabetes in relation to height. The following three models were considered: Model 1 included only height, Model 2 included height as well as demographic and socioeconomic variables, and Model 3 included body mass index in addition to the covariates from Model 2. RESULTS: Overall crude diabetes prevalence was 6.1% (95% CI: 5.0-7.2). Being taller was inversely related with diabetes in women but not in men. In Model 3, a 1-cm increase in women's height was associated with 4% lower odds of having diabetes (OR, 0.96; 95% CI: 0.94-0.99; P = 0.023). CONCLUSIONS: Height is associated with a large reduction in diabetes in women but not in men in Namibia. Interventions that allow women to reach their full growth potential may help prevent the growing diabetes burden in the region.