Scientific communication and scholarly publishing more generally are evolving rapidly under the pressure from funding agencies, regulatory agencies, publishers and the general public for removing the obstacle to data access and to facilitate assessment. This requires having data communication standards in place but that alone is not enough. In this short presentation, we will provide a overview of the landscape of resources relevant to clinical research as well as the latest progress in the field of functional genomics data standards, showing how resources such as Biosharing.org, CDISC standards and ISA format can be harnessed. Following an outline of the key issues, we will share the experience and the lessons learned as part of the IMI eTRIKS initiative the standardization and curation activities involved to improve and facilitate translational research. We will also discuss the opportunities for collaboration and cooperation with other major initiative worldwide, such as the NIH big data to knowledge initiative (NIH BD2K).

eTRIKS (European Translational Research, Informatics and Knowledge Management Services) is a public private partnership made up of 17 pharmaceutical companies, academic institutions and SMEs, jointly financed by the Pharma partners and the Innovative Medicines Initiative of the EU (IMI-JU), with expertise in data management, systems biology, biomedical curation, collaboration and data exchange standards. The goal of the consortium is to leverage the open source transMART platform to provide a series of services to Translational Research and Biomarker research programs, enabling disease stratification and biomarker discovery by:

• Driving the adoption of a common open source platform
• Promoting multi-study data harmonisation
• Developing best practice guidelines and resources for the re-use of research data
• Providing advice and support for translational research projects

In this presentation, we shall describe and outline the current state of the eTRIKS project, illustrate how we are working to support ongoing research projects and how we are providing a hub for a growing ecosystem of open source informatics technologies and providers to support the translational research community.

Very recently, conspicuous effort has been dedicated to describing and predicting the three-dimensional organization of chromosomes inside the eukaryotic nucleus by generic polymer models [1]. In my talk, I will discuss recent results showing that chromosome structure and dynamics can be quantitatively described by a polymer model of decondensing chromosomes which takes into account only minimal physical ingredients like density, stiffness and topology conservation of the chromatin fiber [2-4]. Then, I will present preliminary results concerning how this model can be (1) employed in order to investigate the origin of the visco-elastic properties of the nucleus [5], and (2) suitably generalized for studying the consequences on chromosome structure arising from a mixed composition (10nm-fibers vs. 30nm-fibers) of the underlying chromatin fiber [6].

References:

1. A. Rosa and C. Zimmer, Int. Rev. Cell & Mol. Biol. 307, 275 (2014).
2. A. Rosa and R. Everaers, Plos Comput. Biol. 4, e1000153 (2008).
3. A. Rosa et al., Biophys. J. 98, 2410 (2010).
4. A. Rosa and R. Everaers, Phys. Rev. Lett. 112, 118302 (2014).
5. M. Valet and A. Rosa, in preparation.
6. A.-M. Florescu and A. Rosa, in preparation.

Recent papers have investigated the link between replication timing, replication domains and topologically associated domains. Graph theory, machine learning and signal processing are key examples for methods applied for segmenting the genome based on contact frequency profiles, and for further linking it with replication domains. This presentation will discuss the latest algorithmic advancements and challenges in genome segmentation in the context of investigating the link between replication dynamics and chromatin folding.

An important concept in biology is the link between structure and function. For example, 'sequence makes structure makes function' is a key idea in protein folding. This presentation will address the concept of 'geometry makes function, function makes geometry'. It will present the latest methods for data integration of chromatin conformation and multi omic data, highlighting key results and open challenges.

Restraint-based modeling of genomes has been recently explored with the advent of Chromosome Conformation Capture (3C-based) experiments. We previously developed a reconstruction method to resolve the 3D architecture of both prokaryotic and eukaryotic genomes using 3C-based data. These models were congruent with fluorescent imaging validation. However, the limits of such methods have not systematically been assessed. Here we propose the first evaluation of a mean field restraint-based reconstruction of genomes by considering diverse chromosome architectures and different levels of data noise and structural variability. The results show that: first, current scoring functions for 3D reconstruction correlate with the accuracy of the models; second, reconstructed models are robust to noise but sensitive to structural variability; third, the local structure organization of genomes, such as Topologically Associating Domains, results in more accurate models; fourth, to a certain extent, the models capture the intrinsic structural variability in the input matrices; and fifth, the accuracy of the models can be a priori predicted by analyzing the properties of the interaction matrices. In summary, our work provides a systematic analysis of the limitations of a mean-field restrain-based method, which could be taken into consideration in further development of methods as well as their applications.

3C, 4C and more recently 5C and Hi-C data, were shown to be important for classification of disease taxonomy (for example, in leukaemia).To date, however, there is little established methodology for automated classification and inference. Thus, while imaging data of nuclear architecture, such as Fluorescent In Situ Hybridization (FISH) are commonly used for clinical applications, chromosome conformation data are still far from being adopted. This presentation will discuss some of the computational and bioinformatics challenges involved in data integration and data mining of chromosome conformation data and in their spatial interpretation, towards potential avenues for clinical applications.

An opportunity to address questions and open discussion with Special Session speakers.

Benchmarking is needed for tool assessment and improvement but is complicated by a lack of gold standards, by extensive resource requirements and by difficulties in sharing personal genomic information. To resolve these issues, we launched the ICGC-TCGA DREAM Somatic Mutation Calling Challenge. The BAMSurgeon tool for simulating cancer genomes, and the results of 248 single nucleotide variant analyses of three in silico tumors created with it, will be presented. Different algorithms exhibit characteristic error profiles, and, intriguingly, false positives show a trinucleotide profile very similar to one found in human tumors. Although the three simulated tumors differ in sequence contamination (deviation from normal cell sequence) and in subclonality, an ensemble of pipelines outperforms the best individual pipeline in all cases. BAMSurgeon is available at https://github.com/adamewing/bamsurgeon/.

The ICGC-TCGA DREAM Somatic Mutation Calling Challenge has highlighted the many difficulties in benchmarking and scoring structural variant detection, and has revealed areas of improving detection algorithms. The Challenge includes 206 structural variant analyses of three in silico tumors (created with BAMSurgeon). Different approaches to scoring these analyses for accuracy, and to defining ensembles of these analyses, will be presented. While different structural variant detection algorithms exhibit characteristic error profiles, errors generally tend to be associated with low mapping quality at the predicted breakpoints.

Structural variation (SV) is a major source of genomic variation and plays an important role in cancer genome evolution. However, due to the methodological limitations in aligning and interpreting short reads spanning breakpoints, current methods cannot achieve a high sensitivity and precision. Here, we present a novel algorithm, novoBreak, which comprehensively characterizes a variety of structural breakpoints at base-pair resolution. novoBreak first chops tumor reads into k-mers and indexes them; then by filtering against reference and normal reads, it derives tumor-specific k-mers (novo-kmers); next it clusters reads with the same breakpoints based on the novo-kmers and then locally assembles the reads associated with each breakpoint into contigs. After aligning the contigs to the reference, novoBreak identifies the precise breakpoints and infers various types of SVs. novoBreak consistently performed best in the SV breakpoint calling subchallenges in the ICGC-TCGA DREAM 8.5 Somatic Mutation Calling challenge. The framework of novoBreak can also be applied to discover germline events, gene fusions in RNA-seq data and SV breakpoints in whole exome data. The wider application of novoBreak is expected to reveal comprehensive structural landscape that can be linked to novel mechanistic signatures in cancer genomes.