Data management, bioinformatics and statistics

Introduction

The INF project is the central data management and analysis project of miTarget and covers three essential areas: data management, bioinformatic pipeline development, and statistics.

WP1 – Data management: We established a dedicated research data management (RDM) platform called miTarget-medfdm at the IKMB (https://yoda-mitarget.ikmb.medicsh.de/; accessible via UKSH intranet). In Phase II, among other things, we will expand data management functionalities in terms of automated metadata information retrieval, high-performance (web)calculations and synchronisation, and introduce new training capabilities.

WP2 – Bioinformatic pipelines: In TOFU-MAaPO (https://github.com/ikmb/TOFU-MAaPO), we have implemented state-of-the-art bioinformatics methods for metagenome analysis from raw data analysis to secondary analysis via Nextflow and Singularity/Docker technology to deliver publication-ready results. This is particularly useful for analysing the KINDRED cohort (P1), in which we have metagenomic data from IBD patients (and unrelated healthy controls) and their high-risk relatives, which show a high concordance of genetic and environmental factors.

WP3 – Privacy compliant, ultra-fast bioinformatic web services: Recently, we have implemented ultra-fast bioinformatics web services (https://hybridcomputing.ikmb.uni-kiel.de) using hardware accelerator cards such as FPGAs (Field Programmable Gate Arrays) or GPUs (Graphics Processing Units) for genomewide phasing and imputation (Wienbrandt and Ellinghaus, Bioinform Adv. 2024), statistical interaction (Wienbrandt and Ellinghaus, bioRxiv 2023) as well as peptidome analyses (together with P1 (ElAbd, bioRxiv 2023), that enable secure, energy-saving and ultra-fast processing of genetic-molecular data on privacy-protected computers using two factor authentication (2FA) and other IT security techniques. In Phase II, we will implement the most frequently used Nextflow applications from TOFU-MAaPO (WP2) and “dark matter” analysis (WP2) into a new webservice application named TOFU-MAaPOWeb.

WP4 – Network analysis: We will further develop and extend, evaluate, and apply statistical methods for network-analysis of the microbiome and multi-omics data sets, the latter using microbiome, transcriptome, genome and metabolome data from the same individuals. Furthermore, we plan to modify the concept of a core microbiome using a definition which is not based on individual species abundance or prevalence, but rather network-based (Sharma, Front Genet 2023).

WP5 – Statistical methods for interventional trials: Wewill focus on statistical methods for the identification of predictive biomarkers in randomized clinical trials (RCTs), to ultimately translate microbiome-disease associations into clinical practice. In particular, this will include methods to identify treatment-predictive biomarkers and biomarker-defined subgroups of patients.

WP6 – Statistical support and collaborations: The INF project will be available for statistical support of all other projects within this RU. In particular, we will provide methodological expertise for clinical and pre-clinical interventional studies regarding study design and statistical analysis, e.g., for studies planned in P1.

Aims

We here aim

• To manage an overall data management platform (iRODS) for metagenome data to keep all data organized, documented and searchable
• To develop scalable and reproducible bioinformatic software pipelines (Nextflow) for metagenome analyses
• To customize statistical methods to address project-specific research questions related to the microbiome as a therapeutic target arising from the individual subprojects.