Intersecting Pathways in Bioinformatics and Translational Informatics: A One Health Perspective on Key Contributions and Future Directions

Summary

Objectives: To identify and summarize the top bioinformatics and translational informatics (BTI) papers published in 2022 for the International Medical Informatics Association (IMIA) Yearbook 2023.

Methods: We conducted a comprehensive literature search to identify the top BTI papers, resulting in a set of ten candidate papers. The candidates were reviewed by the section co-editors and external reviewers to select the top three papers from 2022.

Results: From a total of 558 papers, we identified a final candidate list of ten BTI papers for peer-review. These papers apply new statistical frameworks and experimental designs to better capture individual variability in disease and incorporate data that captures differences between single cells and across environmental exposures. In addition, they highlight the importance of model generalization across diverse cohorts and scalability to large medical centers.

Conclusions: We note several important trends in the candidate top BTI papers this year, including a continued focus on developing accurate and scalable computational models to predict disease risk across diverse cohorts and new strategies to capture the molecular heterogeneity of disease.

Keywords

Bioinformatics - genomics - transcriptomics - machine learning - precision medicine

1 Introduction

The year 2022 found us in a crucial moment of reckoning with our interdependence with the environment and the myriad creatures that share our planet. As such, there is an opportunity to integrate One Health into the Bioinformatics and Translational Informatics (BTI) domain. One Health is an approach that aligns with the realization that the health of humans is intrinsically connected to the health of animals and our shared environment.

Work on pursuing the One Health framework, with an emphasis on the interconnectedness of human, animal, and environmental health, has occurred for a few years now. Trinh et al. expanded the One Health approach to include entire microbial communities, showing that environmental and animal microbiomes can influence human health [[1]]. Timme et al. highlighted the necessity of interoperable and open data platforms to facilitate easy sharing and building upon findings in One Health [[2]]. Lustgarten et al. presented a review on the state and potential of veterinary informatics, shedding light on the vital role it plays in bridging veterinary medicine, human medicine, and One Health initiatives despite current limitations [[3]]. These significant contributions underscore the continued development in bioinformatics and translational informatics focusing on an integrative One Health perspective.

For the 2022 BTI best papers, our focus shifted to this One Health theme. In the following synopsis, we will describe our methodology for selecting the top BTI papers of the year that have made significant contributions to the literature this year, including those that improve our understanding of the One Health concept. We will also discuss three important trends highlighted by these papers. We hope that these selections will stimulate further discussion and research into this critical area of health and well-being.

2 Methods

We conducted a comprehensive literature search to identify candidates for the top bioinformatics and translational informatics (BTI) synopsis. To identify articles specifically from 2022, we restricted the query to only return articles with electronic publication dates between January 1, 2022, and December 31, 2022. In addition, we focused the query on the most relevant journals for BTI: Journal of the American Medical Informatics Association (JAMIA), Journal of Biomedical Informatics (JBI), PLoS Computational Biology, Bioinformatics, BMC Bioinformatics, BMC Systems Biology, Nature, Nature Genetics, Nature Biotechnology, Nature Methods, Science, Science Translational Medicine, Clinical Pharmacology and Therapeutics, New England Journal of Medicine, Journal of the American Medical Association (JAMA), Lancet, PLoS Genetics, and Cell.

Using the constraints above, we queried PubMed using the following Medical Subject Headings (MeSH) terms and their derivatives: “bioinformatics”, “computational biology”, “translational research, biomedical”, “genetics, medical”, “genomics”, “gene expression”, “transcriptomics”, “proteomics”, “epigenomics”, “metagenomics”, “omics”, “precision medicine”, “algorithms”, and “machine learning”. All results were identified by a combination of at least one of the following computation-related terms (e.g., “bioinformatics”, “machine learning”) and at least one of the biology-related terms (e.g., “omics”). Finally, to ensure the translational and clinical relevance of our results, we required all articles to be described by the “humans” MeSH term. Our query returned 558 articles for the initial round of editorial review.

In the initial round of editorial review, we read the title and abstract of each article, retaining the ones that were well-aligned with the field of BTI. We evaluated each article that passed the abstract review based on three criteria: novelty, significance, and quality. Novelty refers to originality of the research question or research methodology, significance refers to the potential impact of the paper on BTI, and quality refers to the accuracy of the technical content and interpretation of the results. The editorial review process resulted in 35 candidate articles. We scored these candidate articles in a second round of editorial review based on the above criteria scores and ranked the candidates. This process resulted in a final set of ten candidate articles that were uploaded for external review by domain experts and the other IMIA editors. Each of the ten candidate articles was reviewed in full by at least two external reviewers, both section co-editors, and senior editors. We selected the three top-scoring articles as the best BTI papers of 2022 ([Table 1]). A content summary of each of these best papers can be found in the appendix of this synopsis.

Table 1 Selection of best papers for the 2023 IMIA Yearbook of Medical Informatics for the Bioinformatics and Translational Informatics section. The articles are listed in alphabetical order by the first author's surname.

3 Trends

The ten candidate best papers for 2022 illustrate recent efforts towards improving the representation of individual and molecular variability in models of disease and exemplify previous trends in BTI such as computational models for disease prevention. We provide a brief discussion of the main themes we discovered while evaluating the top BTI papers. These themes highlight important areas of research in this field where we expect to see continued growth in the coming years.

3.1 Applying Proactive Models of Risk to Prevent Disease and Enhance Patient Outcomes

The development and clinical implementation of risk prediction models continues to be a focal point of BTI research. These models apply advanced machine learning techniques to assess potential risk factors and disease subtypes to provide clinically actionable insights.

Grazioli et al. [[4]] propose a Multimodal Variational Information Bottleneck (MVIB) to integrate heterogeneous data modalities into a single predictive framework. The joint encoding leveraged information about microbial species abundance, microbial strain-level markers, and metabolites to generate a predictive model. This approach could potentially offer a proactive model of risk prevention by enabling early disease prediction, thereby contributing to improved patient outcomes. In particular, the incorporation of microbiome data to improve disease prediction for a range of common diseases demonstrates the importance of using One Health concepts to understand human health. Moving beyond a strictly human-centric approach has the potential to improve the impact of precision health initiatives. The survey paper in the 2022 IMIA Yearbook for the BTI section written by Cooper et al. [[7]] explores the interaction between the microbiome and nutrition in more detail.

The work by Wang et al. [[8]] enhances our understanding of the genetic factors that influence physical activity and sedentary behavior; importantly, their study provides insights that may contribute to the development of personalized interventions to promote physical activity and reduce sedentary behavior. In turn, this can help prevent disease and enhance patient outcomes. This is particularly relevant as it recognizes the interconnection between the health of individuals and their environments.

The last three papers in this section build upon a broad foundation of disease risk prediction models in the literature. Of note, each of these papers include a discussion about the performance of such approaches when applied to diverse human populations. In Wang et al. [[9]], the authors demonstrate the potential of genomic data in personalized disease risk assessment, specifically for cancer. By improving our understanding of genetic risk factors and their variability among different populations, healthcare providers can better predict an individual's risk of developing cancer and take proactive measures to prevent disease progression. However, the findings also highlight the importance of inclusivity and diversity in genetic research to ensure the benefits of such risk models can be realized by all population groups. Mishra et al. [[10]] conduct a large cross-ancestry genome-wide association study (GWAS) for stroke, and use their results to identify new risk loci and potential drug targets. Ultimately, the authors generate high performing cross-ancestry and ancestry-specific polygenic risk scores that could improve predictions for individuals with diverse ancestries. Continuing with this theme, Wang et al. [[11]] conclude that implementing preemptive and sequence-based pharmacogenomics prescribing is almost universally applicable, providing a proactive model for disease prevention and enhancing patient outcomes. Additionally, it showcased the efficiency of using health care resources, a critical factor in the scalability of such models in a clinical setting. This proactive approach enables personalized medicine and a more efficient use of health care resources.

3.2 Capturing the Environmental and Molecular Heterogeneity of Disease

Building on the availability of increasingly large datasets and greater emphasis on single-cell genomics assays, many of this year's BTI papers focus on quantifying the environmental and molecular heterogeneity of disease. By capturing and cataloging this variability, we will be able to increase the precision of future research.

For example, Kuppe et al. [[5]] generated a comprehensive spatial and temporal map of cardiac cell types in myocardial infarction patients and controls. This approach recognizes that the disease process in myocardial infarction is not homogeneous; instead, it involves complex interactions between different cell types and their environment at different stages of the disease. The temporal and single-cell approaches allowed the authors to capture the diversity of cellular responses within the heart tissue during disease development and progression. By identifying sets of differentially expressed genes in cells that mark the border between injured and uninjured tissues, the authors characterized the unique profiles of remodeled versus functional myocardium. This provides insights into the heterogeneity of the disease process, reflecting the distinct molecular responses of cells to the injury.

Similarly, Robbe et al. [[12]] generated high resolution maps of copy-number alterations and other genetic variation from DNA sequences of 485 patients with chronic lymphocytic leukemia. This catalog of variation captures distinct patient subgroups within the dataset that associate with disease outcome and differential treatment responses. Understanding the heterogeneity within a single diagnostic category will improve patient outcomes by allowing for more precise treatment options.

There is also environmental heterogeneity for individuals that is not well-captured by traditional population-based estimates. Howe et al. [[13]] sought to address this limitation in GWAS by comparing the results from a large family-based GWAS to existing population-based estimates. They considered a range of behavioral, anthropometric, and molecular traits for their meta-analysis, finding that many phenotypes had smaller GWAS estimates in the family-based analyses. The differences in GWAS estimates were greater for anthropometric and behavioral phenotypes than for the molecular phenotypes (e.g., cholesterol levels), suggesting that leveraging different GWAS structures can better estimate associations for phenotypes influenced by demography and indirect effects. Future insights will require continued data collection for both families and unrelated individuals, as well as the application of multiple GWAS frameworks to compute accurate estimates.

3.3 Considering the Scalability of Computational Models in Clinical Settings

A final theme from the BTI candidate papers this year is an emphasis on the scalability of computational and risk prediction models for use in clinical settings. Some papers, including Wang et al. [[9]] and Wang et al. [[11]], evaluate the deployment of precision models at scale in existing academic medical centers. Others, such as Weitz et al. [[6]] and Tran et al. [[14]], develop models that can more efficiently leverage available data to produce accurate results.

For example, the convolutional neural network (CNN) trained by Weitz et al. [[6]] predicts gene expression directly from hematoxylin and eosin-stained whole slide images in samples from patients with prostate cancer. The final CNN predicts gene expression levels for clusters of co-expressed genes to improve performance and can be used to generate tumor scores normally derived from RNA-sequencing. This showcases a scalable solution that allows for the quantification of gene expression phenotypes directly from imaging, which is particularly beneficial in settings where full molecular phenotyping would be otherwise unattainable. This scalability is crucial for the broad application of such models in clinical settings, thereby facilitating more personalized and effective patient care.

Tran et al. [[14]] apply semi-supervised machine learning approaches to classify central nervous system tumors into known subclasses based on DNA methylation patterns. This allows the authors to leverage a much larger collection of unlabeled methylation data to generate pseudo-labels that can improve classification performance without requiring the expensive clinical resources required to generate fully labeled data. They find that semi-supervised learning accurately classifies tumor classes and can be used to boost performance in supervised learning models. In the future, this approach could be applied to other clinical classification problems where labeled data are limited or to leverage additional public datasets to improve performance.

4 Conclusion

Our efforts to identify the top BTI papers for 2022 have revealed a sustained focus on the development of accurate and scalable computational models for disease risk prediction across diverse cohorts, as well as novel approaches for capturing the molecular heterogeneity of disease. These papers embody the ongoing trend in BTI research towards increasing the representation of individual and molecular variability in disease models and leveraging computational models for proactive disease prevention. We saw work where the utility of advanced machine learning techniques in assessing potential risk factors and disease subtypes in an effort generate clinically actionable insights [[4], [8] [9] [10]]. Moreover, they exemplify the importance of inclusive genetic research and the potential of pharmacogenomics in personalized medicine.

Other papers [[5], [12], [13]] accentuated the importance of capturing the environmental and molecular heterogeneity of disease. These works pave the way for more precise future research by quantifying and cataloging disease variability through single-cell genomics assays, high-resolution genetic variation maps, and different GWAS structures.

Lastly, the scalability of computational models for use in clinical settings is a key theme in the papers [[6], [9], [14]], which illustrate the potential for implementing precision models at scale in existing academic medical centers and how machine learning models can efficiently leverage available data to produce accurate results, facilitating more personalized and effective patient care.

These overarching themes seen across our selection of candidate papers exemplify the continued progress in this field. With such robust foundations in place, we look forward to witnessing further breakthroughs and innovative advancements in these areas in the future.

Die Autoren geben an, dass kein Interessenkonflikt besteht.

Acknowledgements

We are grateful to Martina Hutter and the other IMIA Yearbook editors for their support. We would also like to thank the peer reviewers for their participation in the selection of the top papers for the IMIA Yearbook.

• References

• 1 Trinh P, Zaneveld JR, Safranek S, Rabinowitz PM. One health relationships between human, animal, and environmental microbiomes: a mini-review. One Health Relationships Between Human, Animal, and Environmental Microbiomes: A Mini-Review. Front Public Health 2018 Aug 30;6:235. doi: 10.3389/fpubh.2018.00235.
• 2 Timme RE, Wolfgang WJ, Balkey M, Gubbala Venkata SL, Randolph R, et al. Optimizing open data to support one health: best practices to ensure interoperability of genomic data from bacterial pathogens. One Health Outlook 2020;2(1):20. doi: 10.1186/s42522-020-00026-3.
• 3 Lustgarten JL, Zehnder A, Shipman W, Gancher E, Webb TL. Veterinary informatics: forging the future between veterinary medicine, human medicine, and One Health initiatives-a joint paper by the Association for Veterinary Informatics (AVI) and the CTSA One Health Alliance (COHA). JAMIA Open 2020 Apr 11;3(2):306-17. doi: 10.1093/jamiaopen/ooaa005.
• 4 Grazioli F, Siarheyeu R, Alqassem I, Henschel A, Pileggi G, Meiser A. Microbiome-based disease prediction with multimodal variational information bottlenecks. PLoS Comput Biol 2022 Apr 11;18(4):e1010050. doi: 10.1371/journal.pcbi.1010050.
• 5 Kuppe C, Ramirez Flores RO, Li Z, Hayat S, Levinson RT, Liao X, et al. Spatial multi-omic map of human myocardial infarction. Nature 2022 Aug;608(7924):766-77. doi: 10.1038/s41586-022-05060-x.
• 6 Weitz P, Wang Y, Kartasalo K, Egevad L, Lindberg J, Grönberg H, et al. Transcriptome-wide prediction of prostate cancer gene expression from histopathology images using co-expression-based convolutional neural networks. Bioinformatics 2022 Jun 27;38(13):3462-9. doi: 10.1093/bioinformatics/btac343.
• 7 Cooper K, Clarke M, Clayton JB. Informatics for your Gut: at the Interface of Nutrition, the Microbiome, and Technology. Yearb Med Inform 2023 Jul 6:89-98. doi: 10.1055/s-0043-1768723.
• 8 Wang Z, Emmerich A, Pillon NJ, Moore T, Hemerich D, Cornelis MC, et al. Genome-wide association analyses of physical activity and sedentary behavior provide insights into underlying mechanisms and roles in disease prevention. Nat Genet 2022 Sep;54(9):32-1344. doi: 10.1038/s41588-022-01165-1.
• 9 Wang L, Desai H, Verma SS, Le A, Hausler R, Verma A, et al. Performance of polygenic risk scores for cancer prediction in a racially diverse academic biobank. Genet Med 2022 Mar;24(3):601-609. doi: 10.1016/j.gim.2021.10.015.
• 10 Mishra A, Malik R, Hachiya T, Jürgenson T, Namba S, Posner DC, et al. Stroke genetics informs drug discovery and risk prediction across ancestries. Nature 2022 Nov;611(7934):115-23. doi: 10.1038/s41586-022-05165-3.
• 11 Wang L, Scherer SE, Bielinski SJ, Muzny DM, Jones LA, Black JL, et al. Implementation of preemptive DNA sequence-based pharmacogenomics testing across a large academic medical center: The Mayo-Baylor RIGHT 10K Study. Genet Med 2022 May;24(5):1062-72. doi: 10.1016/j.gim.2022.01.022.
• 12 Robbe P, Ridout KE, Vavoulis DV, Dréau H, Kinnersley B, Denny N, et al. Whole-genome sequencing of chronic lymphocytic leukemia identifies subgroups with distinct biological and clinical features. Nat Genet 2022 Nov;54(11):1675-89. doi: 10.1038/s41588-022-01211-y.
• 13 Howe LJ, Nivard MG, Morris TT, Hansen AF, Rasheed H, Cho Y, et al. Within-sibship genome-wide association analyses decrease bias in estimates of direct genetic effects. Nat Genet 2022 May;54(5):581-92. doi: 10.1038/s41588-022-01062-7.
• 14 Tran QT, Alom MZ, Orr BA. Comprehensive study of semi-supervised learning for DNA methylation-based supervised classification of central nervous system tumors. BMC Bioinformatics 2022 Jun 8;23(1):223. doi: 10.1186/s12859-022-04764-1.

Publikationsverlauf

Artikel online veröffentlicht:
26. Dezember 2023