An assessment of the value of deep neural networks in genetic risk prediction for surgically relevant outcomes.

INTRODUCTION: Postoperative complications affect up to 15% of surgical patients constituting a major part of the overall disease burden in a modern healthcare system. While several surgical risk calculators have been developed, none have so far been shown to decrease the associated mortality and morbidity. Combining deep neural networks and genomics with the already established clinical predictors may hold promise for improvement. METHODS: The UK Biobank was utilized to build linear and deep learning models for the prediction of surgery relevant outcomes. An initial GWAS for the relevant outcomes was initially conducted to select the Single Nucleotide Polymorphisms for inclusion in the models. Model performance was assessed with Receiver Operator Characteristics of the Area Under the Curve and optimum precision and recall. Feature importance was assessed with SHapley Additive exPlanations. RESULTS: Models were generated for atrial fibrillation, venous thromboembolism and pneumonia as genetics only, clinical features only and a combined model. For venous thromboembolism, the ROC-AUCs were 60.1% [59.6%-60.4%], 63.4% [63.2%-63.4%] and 66.6% [66.2%-66.9%] for the linear models and 51.5% [49.4%-53.4%], 63.2% [61.2%-65.0%] and 62.6% [60.7%-64.5%] for the deep learning SNP, clinical and combined models, respectively. For atrial fibrillation, the ROC-AUCs were 60.3% [60.0%-60.4%], 78.7% [78.7%-78.7%] and 80.0% [79.9%-80.0%] for the linear models and 59.4% [58.2%-60.9%], 78.8% [77.8%-79.8%] and 79.8% [78.8%-80.9%] for the deep learning SNP, clinical and combined models, respectively. For pneumonia, the ROC-AUCs were 50.1% [49.6%-50.6%], 69.2% [69.1%-69.2%] and 68.4% [68.0%-68.5%] for the linear models and 51.0% [49.7%-52.4%], 69.7% [.5%-70.8%] and 69.7% [68.6%-70.8%] for the deep learning SNP, clinical and combined models, respectively. CONCLUSION: In this report we presented linear and deep learning predictive models for surgery relevant outcomes. Overall, predictability was similar between linear and deep learning models and inclusion of genetics seemed to improve accuracy.

Adversarial and variational autoencoders improve metagenomic binning.

Assembly of reads from metagenomic samples is a hard problem, often resulting in highly fragmented genome assemblies. Metagenomic binning allows us to reconstruct genomes by re-grouping the sequences by their organism of origin, thus representing a crucial processing step when exploring the biological diversity of metagenomic samples. Here we present Adversarial Autoencoders for Metagenomics Binning (AAMB), an ensemble deep learning approach that integrates sequence co-abundances and tetranucleotide frequencies into a common denoised space that enables precise clustering of sequences into microbial genomes. When benchmarked, AAMB presented similar or better results compared with the state-of-the-art reference-free binner VAMB, reconstructing ~7% more near-complete (NC) genomes across simulated and real data. In addition, genomes reconstructed using AAMB had higher completeness and greater taxonomic diversity compared with VAMB. Finally, we implemented a pipeline Integrating VAMB and AAMB that enabled improved binning, recovering 20% and 29% more simulated and real NC genomes, respectively, compared to VAMB, with moderate additional runtime.

Progressive Cone-Rod Dystrophy and RPE Dysfunction in Mitfmi/+ Mice.

Mutations in the mouse microphthalmia-associated transcription factor (Mitf) gene affect retinal pigment epithelium (RPE) differentiation and development and can lead to hypopigmentation, microphthalmia, deafness, and blindness. For instance, an association has been established between loss-of-function mutations in the mouse Mitf gene and a variety of human retinal diseases, including Waardenburg type 2 and Tietz syndromes. Although there is evidence showing that mice with the homozygous Mitfmi mutation manifest microphthalmia and osteopetrosis, there are limited or no data on the effects of the heterozygous condition in the eye. Mitf mice can therefore be regarded as an important model system for the study of human disease. Thus, we characterized Mitfmi/+ mice at 1, 3, 12, and 18 months old in comparison with age-matched wild-type mice. The light- and dark-adapted electroretinogram (ERG) recordings showed progressive cone-rod dystrophy in Mitfmi/+ mice. The RPE response was reduced in the mutant in all age groups studied. Progressive loss of pigmentation was found in Mitfmi/+ mice. Histological retinal sections revealed evidence of retinal degeneration in Mitfmi/+ mice at older ages. For the first time, we report a mouse model of progressive cone-rod dystrophy and RPE dysfunction with a mutation in the Mitf gene.

Deep integrative models for large-scale human genomics.

Polygenic risk scores (PRSs) are expected to play a critical role in precision medicine. Currently, PRS predictors are generally based on linear models using summary statistics, and more recently individual-level data. However, these predictors mainly capture additive relationships and are limited in data modalities they can use. We developed a deep learning framework (EIR) for PRS prediction which includes a model, genome-local-net (GLN), specifically designed for large-scale genomics data. The framework supports multi-task learning, automatic integration of other clinical and biochemical data, and model explainability. When applied to individual-level data from the UK Biobank, the GLN model demonstrated a competitive performance compared to established neural network architectures, particularly for certain traits, showcasing its potential in modeling complex genetic relationships. Furthermore, the GLN model outperformed linear PRS methods for Type 1 Diabetes, likely due to modeling non-additive genetic effects and epistasis. This was supported by our identification of widespread non-additive genetic effects and epistasis in the context of T1D. Finally, we constructed PRS models that integrated genotype, blood, urine, and anthropometric data and found that this improved performance for 93% of the 290 diseases and disorders considered. EIR is available at https://github.com/arnor-sigurdsson/EIR.