Research

Natural perturbations: Somatic copy number alterations

Natural perturbations like copy number alterations and SNPs can promote cancer development. The development of cancer is a complex process involving the accumulation of multiple mutations and genomic aberrations. A consequence of these alterations is the deregulation of cell signalling pathways central to the control of cell growth, cell fate and other important cellular functions. With our partners at the CRI we aim to characterise disruptions of signalling pathways in tumours and to identify genomic alterations that drive tumour evolution.

Dr Yinyin Yuan (in collaboration with Carlos Caldas' lab) works on statistical methods to quantify the impact of copy-number alterations on gene expression in tumours. The methods she has developed identify differential regulation between breast cancer subtypes by comparing regulatory hotspots (Yuan et al., BIBM 2010; B405). Yinyin also works on image analysis methods for quantitative cellularity scoring of tumour samples, which can greatly increase the signal-to-noise ratio in copy-number profiles (Figure 1; manuscript in preparation).

Figure 1
Automated quantitative cellularity scoring improves detection of copy number alterations. In a stained tumour image (step 1) individual nuclei are classified into different cell types (step 2). Tumour cells cluster together and classification performance is improved by spatial smoothing (step 3). The percentage of tumour cells among all cells yields a quantitative cellularity score, which greatly improves the signal to noise ratio in DNA copy number profiles (step 4).

Dr Roland Schwarz develops statistical methods for sequence analysis based on finite-state transducers. This framework allows for very accurate phylogenetic tree reconstruction (Schwarz et al., PLoS ONE 2010; 5: e15788). We use these methods in collaboration with James Brenton’s lab to infer tumour evolution and its main drivers in ovarian cancer within individual patients.

Dr Mauro Castro, who started in May 2010, develops methods for integrating functional information into diagnostic signatures. While signatures based on individual genes are highly redundant and hard to interpret mechanistically, our assumption is that integrating known pathways and networks in the analysis will lead to easier to interpret signatures.

Experimental perturbations: RNA interference

Experimental perturbations like RNAi are key approaches at the forefront of functional genomics. A goal that is becoming more and more prominent in both experimental as well as in computational research is to leverage gene perturbation screens to the identification of molecular interactions, cellular pathways and regulatory mechanisms. Research focus is shifting from understanding the phenotypes of single proteins to understanding how proteins fulfill their function, what other proteins they interact with and where they act in a pathway. Novel ideas on how to use perturbation screens to uncover cellular wiring diagrams can lead to a better understanding of how cellular networks are deregulated in diseases like cancer.

In our group we work on several projects to analyse gene perturbation screens in terms of pathways and cellular networks. For example, we develop methodology for network analysis of high-throughput RNAi screens (Markowetz, PLoS Comp Biol 2010; 6: e1000655; Wang et al., Bioinformatics 2011; Epub 22 Jan). A particular focus of the lab is on nested effects models (NEMs), a statistical approach that is specifically tailored to reconstruct features of pathways from perturbation effects in downstream reporters (Markowetz et al., Bioinformatics 2007; 23: i305). Based on NEMs, we are developing an integrated experimental and computational approach to identify new branches in the NFκB pathway, a key pathway involved in the immune response as well as in cancer, working in collaboration with the group of Thomas Meyer (Max-Planck Institute for Infection Biology, Berlin).

The theory of NEMs has so far been mainly limited to static snapshots of perturbation effects. In a major conceptual improvement, Xin Wang, a PhD student in the group, has combined hidden Markov models with NEMs to reconstruct rewiring events in pathway topologies from time-series data derived after silencing pathway components. Xin applies his methods to gene expression time-series in mouse embryonic stem cells to infer changes in pathway activity in the early stages of differentiation.

In a related project we are interested in the coordinated interplay between epigenetic, transcriptional, and translational mechanisms that are required for the molecular regulation of stem cell fate. We are approaching this question in collaboration with the group of Ihor Lemischka (Mount Sinai School of Medicine, New York) by undertaking a dynamic systems level study of cell fate changes in murine embryonic stem cells. Global changes in histone acetylation, chromatin-bound RNA polymerase II, mRNA, and nuclear protein levels were measured over five days after down-regulation of Nanog, a key pluripotency regulator (Lu et al., Nature 2009; 462: 358). This data set provides a rich resource that allows us to untangle the complexity of the multilayer regulatory mechanism responsible for stem cell fate. We anchored our analyses on changes in nuclear protein expression and found that many lacked concordant changes in mRNA expression, pointing to important roles for translational and post-translational regulation of ESC fate (Lu et al., Nature 2009; 462: 358). Recently, we complemented these analyses with an in-depth study of the relationship between histone acetylation and gene expression in the same data set (Markowetz et al., PLoS Comp Biol 2010; 6: e1001034).

In the future, our lab plans to strengthen its ties with our experimental collaborators in order to approach pivotal questions in biology and medicine by computationally guided experimentation. Biological and clinical questions motivate the development of novel statistical algorithms, which guide the next round of experiments.