News

On January 15th we have a guest speaker. Olivier Pertz from the University of Bern will talk about his work on Decoding and Re-encoding MAPK Fate Decision Signaling.

Date: January 15th
Time: 4 p.m.
Venue: IRI Life Sciences,
Humboldt-Universität zu Berlin,
Philippstr. 13, Michaelis Building (No 18),
Maud Menten Hall (3rd Floor)
https://goo.gl/maps/9LUWXKXj6pv

Decoding and Re-encoding MAPK Fate Decision Signaling

Olivier Pertz, Institute of Cell Biology, University of Bern
Cells dynamically sense and respond to ever changing external stimuli through sophisticated signaling networks. Accordingly, signaling dynamics rather than steady states control fate decisions. For many signaling pathways, heterogeneous dynamic signaling states occur within distinct cells, explaining fate variability observed within a cell population. Measuring single cell signaling dynamics is therefore key to understand how cellular responses correlate with specific cell fate decisions. Here, we combine biosensor imaging, microfluidics, optogenetics and mathematical modelling to map how different MAPK signalling network circuitries fine tune ERK activity dynamics at the single cell level.
In PC-12 cells, a classic model for fate determination, we observe that different growth factors (GFs) wire the MAPK pathway differently to modify the population distribution of
transient/sustained ERK states that are known to induce proliferation/differentiation fates.
EGF only induces transient ERK states; NGF results in a mix of transient/sustained ERK states, with a sharp increase of the latter at high GF input; FGF gradually modifies the population distribution of transient/sustained ERK states in response to augmenting GF input. The latter feature might be consistent with FGF-dependent fate determination in developmental morphogen gradients. Thus, the distinctive GFs wire the MAPK network differently to control the population distributions of transient/sustained ERK states, and fates. Temporal perturbations applied using microfluidics and optogenetics provide highly informative novel signalling states that allow us to infer the feedback structure inherent to the different networks. Ultimately, this allows us to reprogram fate decisions at will by evoking synthetic dynamic signalling states using temporal perturbations.
I will also discuss on another fate determination system in which MCF10A breast epithelial
cells constantly senses the state of the cell collective, and reacts by spatially tuning  survival and proliferation fates to ensuring a critical cell density necessary for proper barrier function. Here we observe two single-cell ERK signalling modes that consist either of stochastic pulses (in presence of GFs), or of co-ordinated ERK waves across multiple cell layers that originate around apoptotic extruding cells (in absence of GFs or in presence of cytotoxic agents). We show that such ERK activity pulses provide a survival signal for about 3 hours allowing to constantly fine tune fate determination during this epithelial homeostasis process. A further degree of spatio-temporal signalling complexity is observed when these cells are grown as 3D epithelial spheroids which can then explain morphogenetic processes such as lumen formation.
Together, these results showcase the rich variety of information processing/transfer enabled by the MAPK/ERK pathway to warrant robust regulation fate decisions at biologically relevant time/length scales.

On December 19th we have a guest speaker: Donate Weghorn from the Centre of Genomic Regulation, Barcelona will talk about Probabilistic approaches to inference of mutation rate and selection in cancer.

Date: Thursday, December 19th

Time: 4 p.m.

Venue: IRI Life Sciences, Humboldt-Universität zu Berlin,

Philippstr. 13, Michaelis Building (No 18),

Maud Menten Hall (3rd Floor)

https://goo.gl/maps/9LUWXKXj6pv

Probabilistic approaches to inference of mutation rate and selection in cancer

Cancer is a highly complex system that evolves asexually under high mutation rates and strong selective pressures. Cancer genomics efforts have identified genes and regulatory elements driving cancer development and neoplastic progression. The detection of both significantly mutated (positive selection) and undermutated (negative selection) genes is completely confounded by the genomic heterogeneity of the cancer mutation rate. Here, I present an approach to address mutation rate heterogeneity in order to increase the power and accuracy of selection inference. Using a hierarchical model, we infer the distribution of mutation rates across genes that underlies the observed distribution of the synonymous mutation count within a given cancer type. This enables the inference of the probability of nonsynonymous mutations without additional parameters, however explicitly taking into account cancer-type-specific mutational signatures, which are known to be highly distinct. We then augmented our test through integrating information at the single-nucleotide level. Based on a model that accounts for the extended sequence context (> 5-mers) around mutated sites, this second component of the test identifies genes with an excess of mutations in specific nucleotide contexts, which deviate from the characteristic context around neutrally evolving passenger mutations. Using the combined test, we discovered a catalogue of well-known cancer driver genes as well as a long tail of novel candidate cancer genes with mutation frequencies as low as 1% and functional supporting evidence.

We are organizing an upcoming workshop on the highly relevant topic of Deep Learning (DL). Established DL domain experts Dagmar Kainmüller, Jonathan Ronen and Grégroire Montavon will cover the specialized domains of image-classification, deep learning for genomic data integration with applications in cancer diagnosis and interpretation of DL results.

The workshop takes place on January 16th/17th, 2019. More information are available at

https://www.comp-cancer.de/dl_workshop/

Places are limited to 20 and CompCancer graduate school members have prioritized access, however, vacant seats will be given to external participants.

On December 3rd we have a guest speaker: Andreas Deutsch from TU Dresden will talk about The mathematics of brain tumour invasion and progression.

Date: Tuesday, December 3rd

Time: 3 p.m.

Venue: IRI Life Sciences, Humboldt-Universität zu Berlin,

Philippstr. 13, Michaelis Building (No 18),

Maud Menten Hall (3rd Floor)

https://goo.gl/maps/9LUWXKXj6pv

Abstract:

The mathematics of brain tumour invasion and progression

Andreas Deutsch
TU Dresden, Centre for Information Services and High Performance Computing, Dresden, Germany

Tumour invasion and progression may be viewed as collective phenomena emerging from the interplay of individual biological cells with their environment. Cell-based mathematical models as cellular automata, Moran processes, and Markov chains can be used to decipher the rules of interaction. In such models cells are regarded as separate units.

Here, we focus on the dynamics of glioma, which are among the most malignant brain tumours. We introduce lattice-gas cellular automaton models to analyse glioma invasion, define spatial and non-spatial Moran processes based on tumour-specific progression patterns to shed light on the size of the tumour originating niche, and adopt Markov chain models to investigate the origin of heterogeneity in glioblastoma.

References:
1. A. Deutsch, S. Dormann: Cellular automaton modeling of biological pattern formation: characterization, applications, and analysis. Birkhauser, Boston, 2018
2. T. Buder, A. Deutsch, B. Klink, A. Voss-Böhme: Patterns of tumor progression predict small and tissue-specific tumor-originating niches. Front. Oncol., 8, 668, 2019
3. A. Dirkse, A. Golebiewska, T. Buder,..., R. Bjerkvig, A. Deutsch, A. Voss-Böhme, S. P. Niclou: Stem cell-associated heterogeneity in Glioblastoma results from intrinsic tumor plasticity shaped by the microenvironment. Nature Commun., 10, 1 1787, 2019

We are recruiting for 14 open Ph.D. positions in our graduate programme in Berlin. Positions are funded for initially 3 years.

We are looking for outstanding candidates with a Masters degree and a computational background that have a strong passion for developing and applying computational methods important for cancer research. You either have a degree in areas of computer sciences, physics or mathematics and you ideally had contact with life sciences, or you have a degree in the life sciences and you have a strong interest in computing or mathematics.

Before you start your application process you need to register via our admission system.

You will be requested to submit the following documents:

    • Application Form (will be provided after registration)
    • Copies of certificates and transcripts of records
    • Motivation letter
    • Two contacts for letters of reference. The referees will be contacted directly from our admission system.
    • Outline of a possible Ph.D. project relevant to the research topics.

Application deadline is February 24th, 2019. You can select up to three potential supervisors you want to work with. Shortlisted candidates will be invited to introduce themselves at a matching and recruitment event in Berlin in April. The starting date of the programme is June 2019.