Methods of estimating polygenic scores (PGSs) from genome-wide association studies are increasingly utilized. However, independent method evaluation is lacking, and method comparisons are often limited. Here, we evaluate polygenic scores derived via seven methods in five biobank studies (totaling about 1.2 million participants) across 16 diseases and quantitative traits, building on a reference-standardized framework. We conducted meta-analyses to quantify the effects of method choice, hyperparameter tuning, method ensembling, and the target biobank on PGS performance. We found that no single method consistently outperformed all others. PGS effect sizes were more variable between biobanks than between methods within biobanks when methods were well tuned. Differences between methods were largest for the two investigated autoimmune diseases, seropositive rheumatoid arthritis and type 1 diabetes. For most methods, cross-validation was more reliable for tuning hyperparameters than automatic tuning (without the use of target data). For a given target phenotype, elastic net models combining PGS across methods (ensemble PGS) tuned in the UK Biobank provided consistent, high, and cross-biobank transferable performance, increasing PGS effect sizes (β coefficients) by a median of 5.0% relative to LDpred2 and MegaPRS (the two best-performing single methods when tuned with cross-validation). Our interactively browsable online-results and open-source workflow prspipe provide a rich resource and reference for the analysis of polygenic scoring methods across biobanks.

Genotype-to-phenotype mapping is an essential problem in the current genomic era. While qualitative case-control predictions have received significant attention, less emphasis has been placed on predicting quantitative phenotypes. This emerging field holds great promise in revealing intricate connections between microbial communities and host health. However, the presence of heterogeneity in microbiome datasets poses a substantial challenge to the accuracy of predictions and undermines the reproducibility of models. To tackle this challenge, we investigated 22 normalization methods that aimed at removing heterogeneity across multiple datasets, conducted a comprehensive review of them, and evaluated their effectiveness in predicting quantitative phenotypes in three simulation scenarios and 31 real datasets. The results indicate that none of these methods demonstrate significant superiority in predicting quantitative phenotypes or attain a noteworthy reduction in Root Mean Squared Error (RMSE) of the predictions. Given the frequent occurrence of batch effects and the satisfactory performance of batch correction methods in predicting datasets affected by these effects, we strongly recommend utilizing batch correction methods as the initial step in predicting quantitative phenotypes. In summary, the performance of normalization methods in predicting metagenomic data remains a dynamic and ongoing research area. Our study contributes to this field by undertaking a comprehensive evaluation of diverse methods and offering valuable insights into their effectiveness in predicting quantitative phenotypes.

The advent of rapid whole-genome sequencing has created new opportunities for computational prediction of antimicrobial resistance (AMR) phenotypes from genomic data. Both rule-based and machine learning (ML) approaches have been explored for this task, but systematic benchmarking is still needed. Here, we evaluated four state-of-the-art ML methods (Kover, PhenotypeSeeker, Seq2Geno2Pheno and Aytan-Aktug), an ML baseline and the rule-based ResFinder by training and testing each of them across 78 species-antibiotic datasets, using a rigorous benchmarking workflow that integrates three evaluation approaches, each paired with three distinct sample splitting methods. Our analysis revealed considerable variation in the performance across techniques and datasets. Whereas ML methods generally excelled for closely related strains, ResFinder excelled for handling divergent genomes. Overall, Kover most frequently ranked top among the ML approaches, followed by PhenotypeSeeker and Seq2Geno2Pheno. AMR phenotypes for antibiotic classes such as macrolides and sulfonamides were predicted with the highest accuracies. The quality of predictions varied substantially across species-antibiotic combinations, particularly for beta-lactams; across species, resistance phenotyping of the beta-lactams compound, aztreonam, amoxicillin/clavulanic acid, cefoxitin, ceftazidime and piperacillin/tazobactam, alongside tetracyclines demonstrated more variable performance than the other benchmarked antibiotics. By organism, Campylobacter jejuni and Enterococcus faecium phenotypes were more robustly predicted than those of Escherichia coli, Staphylococcus aureus, Salmonella enterica, Neisseria gonorrhoeae, Klebsiella pneumoniae, Pseudomonas aeruginosa, Acinetobacter baumannii, Streptococcus pneumoniae and Mycobacterium tuberculosis. In addition, our study provides software recommendations for each species-antibiotic combination. It furthermore highlights the need for optimization for robust clinical applications, particularly for strains that diverge substantially from those used for training.

Although protein sequences encode the information for folding and function, understanding their link is not an easy task. Unluckily, the prediction of how specific amino acids contribute to these features is still considerably impaired. Here, we developed a simple algorithm that finds positions in a protein sequence with potential to modulate the studied quantitative phenotypes. From a few hundred protein sequences, we perform multiple sequence alignments, obtain the per-position pairwise differences for both the sequence and the observed phenotypes, and calculate the correlation between these last two quantities. We tested our methodology with four cases: archaeal Adenylate Kinases and the organisms optimal growth temperatures, microbial rhodopsins and their maximal absorption wavelengths, mammalian myoglobins and their muscular concentration, and inhibition of HIV protease clinical isolates by two different molecules. We found from 3 to 10 positions tightly associated with those phenotypes, depending on the studied case. We showed that these correlations appear using individual positions but an improvement is achieved when the most correlated positions are jointly analyzed. Noteworthy, we performed phenotype predictions using a simple linear model that links per-position divergences and differences in the observed phenotypes. Predictions are comparable to the state-of-art methodologies which, in most of the cases, are far more complex. All of the calculations are obtained at a very low information cost since the only input needed is a multiple sequence alignment of protein sequences with their associated quantitative phenotypes. The diversity of the explored systems makes our work a valuable tool to find sequence determinants of biological activity modulation and to predict various functional features for uncharacterized members of a protein family.

Combining training data from multiple sources increases sample size and reduces confounding, leading to more accurate and less biased machine learning models. In healthcare, however, direct pooling of data is often not allowed by data custodians who are accountable for minimizing the exposure of sensitive information. Federated learning offers a promising solution to this problem by training a model in a decentralized manner thus reducing the risks of data leakage. Although there is increasing utilization of federated learning on clinical data, its efficacy on individual-level genomic data has not been studied. This study lays the groundwork for the adoption of federated learning for genomic data by investigating its applicability in two scenarios: phenotype prediction on the UK Biobank data and ancestry prediction on the 1000 Genomes Project data. We show that federated models trained on data split into independent nodes achieve performance close to centralized models, even in the presence of significant inter-node heterogeneity. Additionally, we investigate how federated model accuracy is affected by communication frequency and suggest approaches to reduce computational complexity or communication costs.

To explore a robust tool for advancing digital breeding practices through an artificial intelligence-driven phenotype prediction expert system, we undertook a thorough analysis of 11 non-linear regression models. Our investigation specifically emphasized the significance of Support Vector Regression (SVR) and SHapley Additive exPlanations (SHAP) in predicting soybean branching. By using branching data (phenotype) of 1918 soybean accessions and 42 k SNP (Single Nucleotide Polymorphism) polymorphic data (genotype), this study systematically compared 11 non-linear regression AI models, including four deep learning models (DBN (deep belief network) regression, ANN (artificial neural network) regression, Autoencoders regression, and MLP (multilayer perceptron) regression) and seven machine learning models (e.g., SVR (support vector regression), XGBoost (eXtreme Gradient Boosting) regression, Random Forest regression, LightGBM regression, GPs (Gaussian processes) regression, Decision Tree regression, and Polynomial regression). After being evaluated by four valuation metrics: R2 (R-squared), MAE (Mean Absolute Error), MSE (Mean Squared Error), and MAPE (Mean Absolute Percentage Error), it was found that the SVR, Polynomial Regression, DBN, and Autoencoder outperformed other models and could obtain a better prediction accuracy when they were used for phenotype prediction. In the assessment of deep learning approaches, we exemplified the SVR model, conducting analyses on feature importance and gene ontology (GO) enrichment to provide comprehensive support. After comprehensively comparing four feature importance algorithms, no notable distinction was observed in the feature importance ranking scores across the four algorithms, namely Variable Ranking, Permutation, SHAP, and Correlation Matrix, but the SHAP value could provide rich information on genes with negative contributions, and SHAP importance was chosen for feature selection. The results of this study offer valuable insights into AI-mediated plant breeding, addressing challenges faced by traditional breeding programs. The method developed has broad applicability in phenotype prediction, minor QTL (quantitative trait loci) mining, and plant smart-breeding systems, contributing significantly to the advancement of AI-based breeding practices and transitioning from experience-based to data-based breeding.

In recent decades, preterm birth (PTB) has become a significant research focus in the healthcare field, as it is a leading cause of neonatal mortality worldwide. Using five independent study cohorts including 1290 vaginal samples from 561 pregnant women who delivered at term (n = 1029) or prematurely (n = 261), we analysed vaginal metagenomics data for precise microbiome structure characterization. Then, a deep neural network (DNN) was trained to predict term birth (TB) and PTB with an accuracy of 84.10% and an area under the receiver operating characteristic curve (AUROC) of 0.875 ± 0.11. During a benchmarking process, we demonstrated that our DL model outperformed seven currently used machine learning algorithms. Finally, our results indicate that overall diversity of the vaginal microbiota should be taken in account to predict PTB and not specific species. This artificial-intelligence based strategy should be highly helpful for clinicians in predicting preterm birth risk, allowing personalized assistance to address various health issues. DeepMPTB is open source and free for academic use. It is licensed under a GNU Affero General Public License 3.0 and is available at https://deepmptb.streamlit.app/ . Source code is available at https://github.com/oschakoory/DeepMPTB and can be easily installed using Docker ( https://www.docker.com/ ).

The recent development of deep learning methods have undoubtedly led to great improvement in various machine learning tasks, especially in prediction tasks. This type of methods have also been adapted to answer various problems in bioinformatics, including automatic genome annotation, artificial genome generation or phenotype prediction. In particular, a specific type of deep learning method, called graph neural network (GNN) has repeatedly been reported as a good candidate to predict phenotypes from gene expression because its ability to embed information on gene regulation or co-expression through the use of a gene network. However, up to date, no complete and reproducible benchmark has ever been performed to analyze the trade-off between cost and benefit of this approach compared to more standard (and simpler) machine learning methods. In this article, we provide such a benchmark, based on clear and comparable policies to evaluate the different methods on several datasets. Our conclusion is that GNN rarely provides a real improvement in prediction performance, especially when compared to the computation effort required by the methods. Our findings on a limited but controlled simulated dataset shows that this could be explained by the limited quality or predictive power of the input biological gene network itself.

Forensic entomological evidence is employed to estimate minimum postmortem interval (PMImin), location, and identification of fly samples or human remains. Traditional forensic DNA analysis (i.e., STR, mitochondrial DNA) has been used for human identification from the larval gut contents. Forensic DNA phenotyping (FDP), predicting human appearance from DNA-based crime scene evidence, has become an established approach in forensic genetics in the past years. In this study, we aimed to recover human DNA from Lucilia sericata (Meigen 1826) (Diptera: Calliphoridae) gut contents and predict the eye and hair color of individuals using the HIrisPlex system. Lucilia sericata larvae and reference blood samples were collected from 30 human volunteers who were under maggot debridement therapy. The human DNA was extracted from the crop contents and quantified. HIrisPlex multiplex analysis was performed using the SNaPshot minisequencing procedure. The HIrisPlex online tool was used to assess the prediction of the eye and hair color of the larval and reference samples. We successfully genotyped 25 out of 30 larval samples, and the most SNP genotypes (87.13%) matched those of reference samples, though some alleles were dropped out, producing partial profiles. The prediction of the eye colors was accurate in 17 out of 25 larval samples, and only one sample was misclassified. Fourteen out of 25 larval samples were correctly predicted for hair color, and eight were misclassified. This study shows that SNP analysis of L. sericata gut contents can be used to predict eye and hair color of a corpse.

Precision medicine models often perform better for populations of European ancestry due to the over-representation of this group in the genomic datasets and large-scale biobanks from which the models are constructed. As a result, prediction models may misrepresent or provide less accurate treatment recommendations for underrepresented populations, contributing to health disparities. This study introduces an adaptable machine learning toolkit that integrates multiple existing methodologies and novel techniques to enhance the prediction accuracy for underrepresented populations in genomic datasets. By leveraging machine learning techniques, including gradient boosting and automated methods, coupled with novel population-conditional re-sampling techniques, our method significantly improves the phenotypic prediction from single nucleotide polymorphism (SNP) data for diverse populations. We evaluate our approach using the UK Biobank, which is composed primarily of British individuals with European ancestry, and a minority representation of groups with Asian and African ancestry. Performance metrics demonstrate substantial improvements in phenotype prediction for underrepresented groups, achieving prediction accuracy comparable to that of the majority group. This approach represents a significant step towards improving prediction accuracy amidst current dataset diversity challenges. By integrating a tailored pipeline, our approach fosters more equitable validity and utility of statistical genetics methods, paving the way for more inclusive models and outcomes.