Summer School: AI for Bio-Medicine

Basics of RNA Biology for Computer Scientist

This lecture is an introduction to RNA biology and will provide an overview of the features, abundance ratios, and functional diversity of RNAs with a focus on human cells. In addition, we will look at what features of RNAs are used to create libraries for RNA sequencing.

Morning Tea

Next Generation Biobanking

"Advances in the “omics” field bring about the need for a high number of good quality samples. Additionally the access to high quality biosamples and related clinical data is a basic prerequisite for advanced biomedical research. Many omics studies take advantage of biobanked samples to meet this need. Most of the laboratory errors occur in the pre-analytical phase. Therefore evidence-based standard operating procedures for the pre-analytical phase as well as markers to distinguish between ‘good’ and ‘bad’ quality samples taking into account the desired downstream analysis are urgently needed. Advances in analytical biochemistry have recently made it possible to obtain global snapshots of metabolism. Especially the combination of different molecular omics techniques shows major differentiations in the metabolic make-up of the human population. Metabolites may subscribe the risk for a certain medical phenotype, the response to a given drug treatment, the reaction to a nutritional intervention, or environmental challenge. This presentation combines data about biobanking procedures and routines in a big professional clinical biobank and combines it with biobanking research with modern high-throughput data."

Lunch

Tutorial: Types of Biomedical Data/ Sequencing Protocols

The following topics will be covered: data types, quality metrics, mapping to the reference genome, quantification on gene/transcript level, normalisation, differential expression analysis and visualisation.

Afternoon Tea

Tutorial: Understanding RNA sequencing

The following topics will be covered: data types, quality metrics, mapping to the reference genome, quantification on gene/transcript level, normalisation, differential expression analysis and visualisation.

Hands on: Basics of RNA sequencing

The following topics will be covered: data types, quality metrics, mapping to the reference genome, quantification on gene/transcript level, normalisation, differential expression analysis and visualisation.

Morning Tea

Keynote: Machine Learning in Gene Regulation: Simple Methods and Complex Data

This presentation will summarize a few projects where machine learning methods are being employed to unravel gene regulatory mechanisms. The challenge lies in the combination of DNA sequence information with functional measurements on the epigenetic state of the chromatin in order to rationalize in which cellular state a gene gets activated. Time permitting, I will include some philosophical remarks on the importance of the choice of positive and, especially, negative training data.

Lunch

Special Session: Compression of DNA Sequencing Data

With the release of the latest generations of sequencing machines, the cost of sequencing a whole human genome has dropped to less than US$1,000. The potential applications in several fields lead to the forecast that the amount of DNA sequencing data will soon surpass the volume of other types of data, such as video data. This development hinders the widespread use of DNA sequencing data in fields such as individualized precision medicine. In this special session we will introduce you to novel data compression technologies that aim at enhancing storage, transmission, and processing of DNA sequencing data. We will show how these new developments support the democratization of DNA sequencing data and help realize its yet undiscovered potential. This special session will consist of introductory talks followed by a hands-on session.

Afternoon Tea

Gut Microbiome Data Analysis: from 16S Gene Sequencing to Complete Metagenomes

"There is an intimate relationship between the gut microbiome and the epithelial cells of the gut; disruptions in the balance of this relationship (dysbiosis) are linked with several diseases, including cancer and immune-metabolic disorders. The gut microbiome comprises the genome of all the microorganisms that inhabit the gut. It plays a crucial role in the body's nutritional, metabolic, physiological and immunological processes. It also produces several metabolites such as short-chain fatty acids (SCFAs), branched-chain amino acids, vitamins, and hormones that affect several metabolic and inflammatory pathways. Therefore shifts in host-microbiome interactions have a clinical impact on health, especially in metabolic and immunological diseases. In the past decade, there has been significant progress using genomic sequencing and computational pipelines to interpret the human gut microbiome, mainly by using 16S rRNA sequencing and shotgun metagenomics. Both tools have advantages and limitations and are the cornerstone for interpreting microbiome state and its relationship with disease. However, interpreting and discussing gut microbiome results can be challenging, especially with a lack of standard parameters and reference data. We will discuss the current evidence in host-microbiome interactions and develop a hands-on session to analyze and visualize microbiome data considering compositionality. "

Pediatric Oncology

Medicine is undergoing a paradigm shift from phenotyping to genotyping. An open modular architectural framework for tools, models, and services together with a legal framework for data privacy and data security is developed in different European projects in cooperation between basic scientists, clinicians, informaticians and modellers as well as ethicist and lawyers. Models using AI for different cancer types are developed to foster translational medicine. One of the cancer types is nephroblastoma, the most common kidney cancer in childhood.The developed model for nephroblastoma is clinically driven to enhance decision support. This is guaranteed by an IT infrastructure that is able to handle and share heterogeneous data sets composed of clinical, imaging and basic research data. The infrastructure enables standardization and semantic data interoperability and can integrate the developed models. A legal framework ensures data protection and privacy by implementation of policies for accessing data, services, tools, and models.

The evolution of treatment in pediatric cancer - from no treatment to individualized approaches

After the end of World War II, the discovery of cytotoxic substances led to development of the first systematic clinical approaches to malignant diseases – the true birth of cancer chemotherapy. In 1948, the report of a 3-months remission in ten of 16 children with leukemia undergoing folate monotherapy documents the first successful chemotherapy for childhood leukemia. In 1956 the first successful treatment of a solid tumor by antifolates was reported. It took another ten years until combination chemotherapy was developed which resulted in the first long-term remissions of pediatric leukemia. During the next decades, continuous optimization of polychemo- and radiation therapy in parallel with improved clinical and biological characterization of pediatric cancers led to more refined risk-adapted treatment stratification. This allowed to better balance the risk of disease recurrence and therapy-related toxic side-effects and resulted in growing survival rates. In more recent years, supported by the enormous technical progress in sequencing technologies, biological characterization at both tumor and host levels progressed towards truly personalized genomics which currently results into “super-stratified” treatment approaches. From now on, the challenge will be to integrate the growing complexity of comprehensive data at different biological levels with clinical data and to decipher those pathways integral to malignant disease and treatment response. Keeping in mind an ever growing number of new substances entering the drug market for malignant diseases, it will be difficult to evaluate the results of integrated efforts in context of numerous new drugs and their combinations in classical clinical trial structures. Therefore, innovative computational and mathematical tools will be needed to allow the transition from “super-stratified” to truly individualized pediatric cancer medicine.

Closing Remarks

Morning Tea and Collaboration Time