Applied Computational Biology
The members of the Applied Computational Biology group at the IEG cover a wide range of scientific disciplines, all possessing deep IT knowledge and experience and a strong command of the IT toolbox.
As a team, we work at the interface between biology and applied computer sciences in our institute as well as in external collaborations.
The members of the Applied Computational Biology group at the IEG cover a wide range of scientific disciplines, all possessing deep IT knowledge and experience and a strong command of the IT toolbox.
As a team, we work at the interface between biology and applied computer sciences in our institute as well as in external collaborations.
About our Work
In our scientific branch, we initiate and take lead in projects that apply artificial intelligence methods on large mouse phenotyping datasets. We use internal data from the German Mouse Clinic (GMC) as well as external data from the International Mouse Phenotyping Consortium (IMPC) and other sources.
Such projects aim to
- uncover so far unknown genome-phenome relationships and to identify new candidate genes for models of human disease
- speed up or even enable data analysis in depths not possible so far
- implement objective, unbiased data processing pipelines for the re-analysis of large phenotyping raw data collections
- provide automated and powerful QC tools for the highly standardised GMC phenotyping pipeline.
We develop and maintain highly customised IT infrastructure and IT solutions for our institute. The German Mouse Clinic runs a large collection of special hardware devices and software to measure and capture phenotyping data. Our in-house developed application "MausDB" is a modular laboratory information management system (LIMS) that integrates work planning as well as capture, management and analysis/visualisation of data, providing a high degree of automation [2,3]. Our team provides high-availability service of MausDB and all connected subsystems, as the GMC runs a clocked, time-critical high-throughput operation.
We are also responsible for management, integration, curation and quality control of GMC data using automated and manual procedures following our internal SOPs.
In a center-wide commitment, we have developed and continuously maintain a generic version of MausDB for animal facility management at >20 HMGU institutes, which enables automated and law-compliant generation of the "Jahresstatistik" according to VersTierMeldV and 2010/63/EU.
We internally support IEG research projects with our IT expertise and our tools. This involves, but is not limited to, performing custom SQL queries on our LIMS to extract specific project data, statistical consulting for experimental design and data analysis as well as development of custom big data processing and data analysis solutions, involving AI methods.
In our scientific branch, we initiate and take lead in projects that apply artificial intelligence methods on large mouse phenotyping datasets. We use internal data from the German Mouse Clinic (GMC) as well as external data from the International Mouse Phenotyping Consortium (IMPC) and other sources.
Such projects aim to
- uncover so far unknown genome-phenome relationships and to identify new candidate genes for models of human disease
- speed up or even enable data analysis in depths not possible so far
- implement objective, unbiased data processing pipelines for the re-analysis of large phenotyping raw data collections
- provide automated and powerful QC tools for the highly standardised GMC phenotyping pipeline.
We develop and maintain highly customised IT infrastructure and IT solutions for our institute. The German Mouse Clinic runs a large collection of special hardware devices and software to measure and capture phenotyping data. Our in-house developed application "MausDB" is a modular laboratory information management system (LIMS) that integrates work planning as well as capture, management and analysis/visualisation of data, providing a high degree of automation [2,3]. Our team provides high-availability service of MausDB and all connected subsystems, as the GMC runs a clocked, time-critical high-throughput operation.
We are also responsible for management, integration, curation and quality control of GMC data using automated and manual procedures following our internal SOPs.
In a center-wide commitment, we have developed and continuously maintain a generic version of MausDB for animal facility management at >20 HMGU institutes, which enables automated and law-compliant generation of the "Jahresstatistik" according to VersTierMeldV and 2010/63/EU.
We internally support IEG research projects with our IT expertise and our tools. This involves, but is not limited to, performing custom SQL queries on our LIMS to extract specific project data, statistical consulting for experimental design and data analysis as well as development of custom big data processing and data analysis solutions, involving AI methods.
Selected Publications of the Group
2023 Mammalian Genome
2023 Mammalian Genome
Echo2Pheno: a deep-learning application to uncover echocardiographic phenotypes in conscious mice
2022 Nucleic Acids Research
CORUM: the comprehensive resource of mammalian protein complexes–2022
2022 BMC Neuroscience 23, 81
2018 Nature Communications 9:288
Identification of genetic elements in metabolism by high-throughput mouse phenotyping
2015 Mammalian Genome 26, 467–481
2008 BMC Bioinformatics 9:169
Topics for Master thesis
In our efforts to gain new insights from large amounts of mouse phenotyping data and to automate data analysis, we are always looking for highly motivated and qualified students to join the group.
Automated analysis of mouse ECG signals using machine learning techniques is the subject of a Master’s thesis. This will be supervised together with the German Mouse Clinic (GMC), where the data was also generated in a high-throughput phenotyping process.
The application of artificial intelligence to electrocardiography is developing rapidly. Sophisticated AI algorithms can already analyse ECG signals from patients, but are lacking for rodents. However, mouse models are important for cardiovascular disease research because they have comparable ECG waveforms (except for J wave in mice) and similar parameter sets. However, a 10-fold increase in heart rate in the mice quickly generates large datasets. Human interpretation requires a high level of expertise. Advanced AI-based methods have enabled rapid, human-like ECG interpretation. In addition, they have discovered signals and patterns that have so far remained largely undetected by human interpreters.
The project aims to:
- generate a mouse-specific AI tool for fully automated analysis of high-throughput ECG signals from the GMC
- compare the manual and the automated evaluation across all data
- confirm new ECG findings with cardiac echocardiography data.
In our efforts to gain new insights from large amounts of mouse phenotyping data and to automate data analysis, we are always looking for highly motivated and qualified students to join the group.
Automated analysis of mouse ECG signals using machine learning techniques is the subject of a Master’s thesis. This will be supervised together with the German Mouse Clinic (GMC), where the data was also generated in a high-throughput phenotyping process.
The application of artificial intelligence to electrocardiography is developing rapidly. Sophisticated AI algorithms can already analyse ECG signals from patients, but are lacking for rodents. However, mouse models are important for cardiovascular disease research because they have comparable ECG waveforms (except for J wave in mice) and similar parameter sets. However, a 10-fold increase in heart rate in the mice quickly generates large datasets. Human interpretation requires a high level of expertise. Advanced AI-based methods have enabled rapid, human-like ECG interpretation. In addition, they have discovered signals and patterns that have so far remained largely undetected by human interpreters.
The project aims to:
- generate a mouse-specific AI tool for fully automated analysis of high-throughput ECG signals from the GMC
- compare the manual and the automated evaluation across all data
- confirm new ECG findings with cardiac echocardiography data.