gms | German Medical Science

Diseases are multifaceted, and their manifestation can be very individual. This is reflected in the trend toward precision medicine. For this individualization of diagnosis and treatment, molecular data plays an increasingly important role. These data include, among others, genomic, transcriptomic, and proteomic data. The measurement and evaluation of these data are treated in the corresponding fields of genomics, transcriptomics, and proteomics. The connection to medicine is made through the area of medical omics, which also considers data integration.

In recent years, we have offered workshops on this topic at various local and international events and the annual meetings of the GMDS. The work in this special issue stems from two events, the 1^st International Applied Bioinformatics Conference and the 67^th Annual Meeting of the GMDS.

Six contributions have been selected after peer review. They reflect the broad spectrum of medical omics, which relates not only to data and their integration but also to the coherence of the data structure, including the connection to patients and knowledge databases.

Beukers and Allmer [1] deal with the possibility of automated analysis of transcriptomic data. Such data form the basis for higher-value studies in medical omics and reflect the cellular state. The authors compare three workflow management systems (Galaxy, KNIME, and CLC) for the creation and execution of RNA-seq data analysis workflows. They find that each tool has strengths and weaknesses and that the workflows created by each tool lead to different results. The authors also provide recommendations on which tool should be used and share their workflows to promote the development of best practices for RNA-seq data analysis.

The next step after data analysis of raw data is the integration with previously acquired knowledge. Unlu Yazıcı and colleagues [2] write that individual omics datasets are insufficient to fully understand the molecular mechanisms of complex diseases that are influenced by multiple factors. They use machine learning to integrate multiple omics data. They use existing data to create models for prediction and classification. The authors further discuss the importance of interpreting the biological significance of model outputs with corresponding biological mechanisms and the potential clinical applications.

A similar approach is presented by Königs and Dietrich [3]. However, other data types are used here, and the prediction is a pathway. Pathway enrichment is used to analyze gene expression data and identify significant pathways, while the PharMe-BINet database uses correction methods to reduce false-positive results. An analysis was performed on gene expression data from T-cell lymphomas.

Once data has been obtained at the molecular level, and knowledge has been integrated, it should find application in the hospital. For this, Raupach [4] proposes an integration method. A software module integrates molecular data into a hospital information system to improve the safety of drug therapy and performs pharmacogenetic reviews in an existing drug therapy workflow. The results show the importance of regular, pre-emptive genotyping of patients and the prevalence of genetic variants in the population.

While creating and testing new drugs (also for precision medicine) is essential, it is equally important to research why studies fail. Friedrichs [5] presents a web-based platform that categorizes the reasons for the failure of clinical studies. As of September 23, 2022, the database contains 14,232 failed studies, of which 51.5% are already commented on. The goal is to use the data for information acquisition for decision-making and the reuse of drugs.

All of the presented works go hand in hand with Savoska and colleagues’ work [6], which combines various medical data in a personalized way. The authors propose a cloud-based model to integrate health and medical data from multiple sources, including electronic health records and measurement sensors, into a personal health record to improve the individualized prediction of diseases and treatment of patients. Omics data should also be integrated into this model, which complies with the necessary data security and data protection standards.

These contributions show the broad spectrum of the field of medical omics but cannot reflect its breadth. Other aspects, such as teaching, application, and quality assurance, have yet to be considered here. This highlights the importance of establishing the field of medical omics as a working group in the GMDS. We encourage all interested parties to contact and work with us to advance this critical endeavor.

The authors declare that they have no competing interests.