Identification and prioritization of candidate causal genomic variations from individuals affected by ASD
Authors: Giovanni Spirito (1), Diego Vozzi (2), Martina Servetti (3), Margherita Lerone (3), Maria Teresa Divizia (3), Giulia Rosti (3), Livia Pisciotta (4), Lino Nobili (4), Irene Serio (4), Stefano Gustincich (2), Remo Sanges (1)
Next generation sequencing (NGS) technologies enabled the extensive study of the genomics underlying human diseases. Namely whole exome sequencing (WES) represents a cost-efficient method which can lead to the detection of multiple classes of genomic variants and the discovery of novel disease-associated genes. One of the drawbacks of this approach however, is the large number of genomic variants detected in each analysis. Automated variant prioritization strategies are therefore required. This is particularly important in the case of complex disease such as ASD, whose genetic etiology is still poorly understood. To this aim we built a custom computational framework capable, from raw WES data, to automatically detect four classes of genomic variants (SNPs, indels, copy number variants and short tandem repeat variants) and prioritize them in regards to their relevance to a specific phenotype. We tested this framework on a selection of 29 trios including probands affected by severe and undiagnosed rare phenotypes and a small cohort of 10 trios all featuring healthy parents and one offspring affected by autism spectrum disorder (ASD). We were able to successfully detect rare and de novo high penetrance variants which have been validated and confirmed as causative among the undiagnosed probands. In the specific case of the ASD cohort we could highlight several genes which are not implicated in autism susceptibility, but nevertheless whose connections to genes relevant for ASD could suggest a possible involvement in the phenotype. Furthermore, our approach enabled us to detect several instances characterized by the presence of multiple candidate variants within genes belonging to the same canonical pathway in one proband. Our workflow allows to detect and prioritize multiple classes of genomic variants in order to both highlight rare high penetrance disease-causative mutation, and possibly reconstruct the genomics at the basis of complex ASD phenotypes.
(1) SISSA, Italy, (2) IIT, Italy, (3) Gaslini Institute, Italy, (4) University of Genova, Italy
Omics data integration for the identification of cell-type-specific gene regulatory networks and regulatory variants in Parkinson’s disease
Authors: Borja Gomez Ramos (1,2), Jochen Ohnmacht (1,2), Nikola de Lange (2), Aurélien Ginolhac (1), Aleksandar Rakovic (5), Christine Klein (5), Roland Krause (2) , Marcel H. Schulz (6), Thomas Sauter (1), Rejko Krüger (2,3,4) and Lasse Sinkkonen (1)
Genome-Wide Association Studies (GWAS) have identified many variants associated with different diseases. However, it is still a challenge to make sense of this data as the majority of genetic variants are located in non-coding regions, complicating the understanding of their functionality. In the last few years, it has been found that non-coding genetic variants concentrate in regulatory regions in the genome, which are cell type and cell-stage specific. In this project, we seek to identify functional Parkinson’s disease GWAS non-coding genetic variants that could make carriers more prone to developing PD. To do so, we are using induced pluripotent stem cell (iPSC) technology to differentiate somatic cells into midbrain dopaminergic (mDA) neurons, astrocytes and microglia. Assessing their chromatin accessibility, active chromatin regions and transcriptome, we can identify crucial regulatory regions in the genome, key transcription factors and derive the gene regulatory networks for the three different cell types. Then, we will map the non-coding genetic variants to the different regulatory regions and predict their effect in silico for the subsequent validation in vitro. This innovative approach will also identify novel factors controlling cell fate and cell identity.
(1) Life Sciences Research Unit, University of Luxembourg, Luxembourg, (2) Luxembourg Centre for Systems Biomedicine (LCSB), University of Luxembourg, Luxembourg, (3) Centre Hospitalier de Luxembourg (CHL), Luxembourg, (4) Luxembourg Institute of Health (LIH), Luxembourg, (5) Institute of Neurogenetics, University of Lübeck, Germany, (6) Institute for Cardiovascular Regeneration, Uniklinikum and Goethe University Frankfurt, Germany
The 4th EMBL Conference: Cancer Genomics (4 – 6 November 2019) brought together over 240 scientists in the field of cancer research to present the latest findings in cancer functional genomics, systems biology, cancer immunogenomics and epigenomics, as well as their translation and clinical impact.
123 posters were presented at the two poster sessions, out of which two were selected as the winners by popular vote.
Infinite sites violations during tumour evolution reveal local mutational determinants
The infinite sites model of molecular evolution requires that every base in the genome is mutated at most once. It is a cornerstone of (tumour) phylogenetic analysis, and is often implied when calling, phasing and interpreting variants or studying the mutational landscape as a whole. It is unclear however, whether this assumption holds in practice for bulk tumour samples. Here we provide frameworks to model and detect infinite sites violations, identifying 24,459 in total, including 6 candidate biallelic driver events, in 700 bulk tumour samples (26.3%) from the ICGC/TCGA Pan-Cancer Analysis of Whole Genomes project. Violations generally occur at mutational hotspots and their frequency and type can accurately be predicted from the overall mutation spectrum. In melanoma, their local sequence context evidences how not only ETS, but also NFAT-family transcription factor binding creates hotspots for UV-induced cyclobutane pyrimidine dimer formation. In colorectal adenocarcinoma, violations reveal hypermutable special cases of the trinucleotide mutational contexts identified in POLE-mutant tumours. Taken together, we reveal the infinite sites model breaks down at the bulk level for a considerable fraction of tumours. These results warrant a careful evaluation of current pipelines relying on the validity of the infinite sites assumption, especially when scaling up to larger sets of mutations and lineages in the future.
A total of 189 posters were presented, from which two were singled out as the winners by popular vote.
Characterization of the genomic and splicing features of long non-coding RNAs using bioinformatics approaches
Authors: Monah Abou Alezz, Ludovica Celli, Giulia Belotti, Silvia Bione, Institute of Molecular Genetics L. L Cavalli-Sforza – National Research Council, Italy
Recent developments in deep sequencing approaches have simulated the continuous discovery of a significantly large number of novel long non-coding RNA (lncRNA) genes loci in the genomes. Long non-coding RNAs are recognized as a new class of regulatory molecules despite very little is known about their functions in the cellular processes. Due to their overall low expression level and tissue-specificity, the identification and annotation of lncRNA genes still remains challenging. The characterization of lncRNAs’ features is crucial to understand and get functional insights on their mechanisms of action. We exploited recent annotations by the GENCODE compendium to characterize the genomic and splicing features of long non-coding genes, in comparison to protein-coding ones, in the human and mouse genome by using bioinformatics approaches. Our analysis highlighted differences between the two classes of genes in terms of gene architecture regarding exons and introns length, GC-content, and the combinatorial patterns of chromatin marks and states. Moreover, significant differences in the splice sites usage were observed between long non-coding and protein-coding genes. While the frequency of non-canonical GC-AG splice junctions represents about 0.8% of total splice sites in protein-coding genes, we identified a remarkable enrichment of the GC-AG splice sites in long non-coding genes, both in human (3.0%) and mouse (1.9%). In addition, we identified peculiar characteristics of the GC-AG introns in terms of donor and acceptor splice sites strength, poly-pyrimidine tract, intron length, and a positional bias of GC-AG junctions being enriched in the first intron. Genes containing at least one GC-AG intron were found conserved in many species across large evolutionary distances, more prone to alternative splicing and a functional analysis pointed toward their enrichment in specific biological processes such as
Authors: Bastian Fromm (1), Diana Domanska (2), Eirik Hoye (3), Vladimir Ovchinnikov (4), Wenjing Kang (5), Ernesto Aparicio-Puerta (6), Morten Johansen (7), Kjersti Flatmark (3), Anthony Mathelier (8), Hovig
Eivind (3), Michael Hackenberg (6), Marc Friedländer (5), Kevin Peterson (9)
Non-coding RNAs (ncRNA) have gained substantial attention due to their roles in human disorders and animal development. microRNAs (miRNAs) are unique within this class as they are the only ncRNAs with individual gene sequences conserved across the animal kingdom. Bona fide miRNAs can be clearly distinguished from the myriad small RNAs generated in cells by a set of unique criteria. Unfortunately, recognition and utilization of these clear and mechanistically well understood features is not a common practice. We addressed this by extensively expanding our curated miRNA gene database MirGeneDB to 45 organisms that represent the breadth of Metazoa. By consistently annotating and naming more than 11,000 miRNA genes in these organisms, we show that previous miRNA annotations contained not only many false positives, but surprisingly many false negatives as well. Indeed, curated miRNA complements of closely related organisms are very similar and can be used to reconstruct evolution of miRNA genes, families and biogenesis across more than 1 billion years of evolution. MirGeneDB represents a robust platform for providing deeper and more significant insights into the biology of miRNAs, possible sources of mis-regulation, and evolutionary mechanisms. MirGeneDB is publicly and freely available under http://mirgenedb.org/.
(1) Science for Life Laboratory, Sweden (2) Department of Informatics, University of Oslo, Oslo, Norway (3) Department of Tumor Biology, Institute for Cancer Research, The Norwegian Radium Hospital, Oslo University Hospital, Oslo, Norway (4) School of Life Sciences, Faculty of Health and Life Sciences, University of Nottingham, United Kingdom (5) Stockholm University, SciLifeLab, Sweden (6) Department of Genetics, Faculty of Sciences, University of Granada, Granada, Spain (7) Institute for Medical Informatics, The Norwegian Radium Hospital, Oslo University Hospital, Oslo, Norway (8) Centre for Molecular Medicine Norway (NCMM), Nordic EMBL Partnership, University of Oslo, Oslo, Norway (9) Department of Biological Sciences, Dartmouth College, Hanover, New Hampshire, United States of America
Mycobacteria have a unique membrane structure with a complex hydrophobic outer-membrane rich in mycolic acids. To transport substances across this impermeable barrier, mycobacteria rely on a highly specialised translocation machinery – the Type VII secretion system (T7SS). Pathogenic mycobacteria encode up to five distinct T7SSs ESX-1 to 5 . Our previous work characterised the structure of the of the inner-membrane complex of the ESX-5 T7SS from Mycobacterium xenopi using negative stain electron microscopy, revealing a hexameric 1.8 MDa complex comprising the four conserved core components: EccB5, EccC5, EccD5 and EccE5 . The large cytosolic domain of EccC5, an FtsK/SpoE-like ATPase, is absent in our current EM map due to its conformational flexibility, which may be required to accommodate a range of protein substrates. Our current work aims to understand the role of EccC5 in secretion. In isolation this component can oligomerise into a hexameric ring-like conformation, as observed for other ATPases in this family. In addition, chemical cross-linking of the ESX-5 complex coupled with mass spectrometry (XL-MS) supports the oligomerisation of EccC5 in the secretion complex, suggesting that it may form a channel or ‘translocation tunnel’. Thus, we propose that EccC5 may exist in two conformational states: an extended, flexible monomeric state and a more compact hexameric state. Using an integrative structural biology approach, we are combining structures of isolated proteins derived from X-ray crystallography and electron microscopy studies with XL-MS data. Together these data aim to further elucidate the secretion pathway across the mycobacterial cell envelope.
 Houben, E. N. G., et al. Take five — Type VII secretion systems of Mycobacteria. Biochim. Biophys. Acta – Mol. Cell Res.1843, 1707–1716 (2014).
 Beckham, K. S. H. et al. Structure of the mycobacterial ESX-5 type VII secretion system membrane complex by single-particle analysis. Nat. Microbiol.2, 17047 (2017).
Emergence of multidrug-resistance poses serious threat to the society. One of the effective way by which bacteria gain drug resistance is through active efflux of antibiotics and other antibacterial compounds using multidrug efflux transporters. Among the battery of efflux pumps present in pathogenic bacteria, our work is focused on QacA, a drug-proton anitiporter (DHA) with 14-transmembrane helices that provide resistance to methicillin resistant Staphylococcus aureus (MRSA) strain, with homologs present in other pathogenic organisms. QacA is a highly promiscuous transporter, capable of effluxing diverse array of monovalent and divalent cationic antibacterial compounds and dyes. This study using a homology model, dissects the role of six protonatable residues present in the transport vestibule of QacA. Systematic mutagenesis resulted in identification of D34 (TM1) and E407 (TM13) as crucial residues and D323 (TM10) and D411 (TM13) as conditional residues needed for transport process of QacA. Whole cells, inside-out vesicles, substrate-induced proton release and microscale thermophoresis based assays were used to investigate the transport and binding properties of the transporter and its mutants. The activity of purified protein was checked with reconstituted QacA in a proteoliposome using substrate-induced proton transport assay. We identify two sites, D34 and D411 playing vital role in recognition of most of the substrates tested while E407 facilitates substrate efflux as a protonation site. It was also observed that E407 has an additional role as a recognition site for the transport of dequalinium, a divalent quaternary ammonium compound. These observations rationalize the promiscuity at the residue level of QacA for diverse substrates. The study identifies the role of acidic residues in QacA with implications for substrate recognition, promiscuity and processive transport in multidrug efflux transporters, related to QacA.
(1) Indian Institute of Science, India, (2) Stockholm University, Sweden
Biophysical analysis of circularized MSP nanodiscs for structural studies
Authors: Melina Daniilidis (1), Ralf Stehle (1), Franz Hagn (1,2)
Structure and dynamics of membrane proteins are crucial aspects for understanding functional properties of this protein class. Unfortunately, stabilizing them in their isolated form is still difficult. By incorporating membrane proteins into nanodiscs, they can be studied in a native-like
environment using biochemical and structural methods. However, thermal and long-term stability of small nanodiscs limit these studies and make it difficult to carry out nuclear magnetic resonance spectroscopy (NMR) measurements at elevated temperatures. Circularized membrane scaffold proteins (MSPs) produced via split-inteins have been shown to be more stable and homogenous than their linear counterparts. However, their biophysical properties, as well as suitability for membrane protein insertion and structural studies have not yet been assessed in a systematic manner. Thus, we examined circular and linear nanodiscs of varying size using several biophysical methods. An important issue for NMR structural studies is that the size and shape of circular nanodiscs do not expand above the phase transition temperature, increasing their homogeneity and reducing their size as compared to linear nanodiscs at high temperatures. 1H,15N-TROSY experiments could demonstrate that circular MSP1D1 nanodiscs with incorporated VDAC-1 are stable at higher temperatures, making it possible to obtain high-resolution NMR spectra of superior quality. Furthermore, NMR relaxation experiments were carried out to compare rotational correlation times of VDAC-1 in circular and linear nanodiscs, respectively. Despite the higher molecular weight, the circular nanodiscs showed lower rotational correlation times, which corroborated the biophysical results on the temperature-dependecy of the nanodisc diameter and homogeneity. The presented data demonstrate that these very stable circularized MSPs are well applicable to the study of membrane proteins in a lipid environment by NMR, but also other structural methods like electron microscopy.
(1) Technical University of Munich, Germany, (2) Helmholtz Zentrum München, Germany
Poster currently not available
Reconstitution of the activity of RND efflux pumps into proteoliposomes
Authors: Dhenesh Puvanendran, Quentin Cece, Martin Picard, IBPC, France
Efflux pumps are the major systems in bacterial resistance against antibiotics. They are classified by the energy needed to be active (ATP hydrolysis or ion counter-transport). Efflux pumps from the RND (Resistance, Nodulation, and cell Division) family use a proton gradient to be active and are composed of three proteins: a membrane fusion protein (MFP) and a transporter (RND) in the inner membrane, and an Outer Membrane Factor (OMF) localized in the outer membrane. We focus on the MexA-MexB-OprM efflux pump from Pseudomonas aeruginosa.The overall goal of my research is to measure in vitro the velocity of transport by efflux pumps. To that end, we reconstitute MexA and MexB as one population of proteoliposome, and OprM as another population of proteoliposome. The whole tripartite pump forms upon association of the respective populations of liposomes. The proof of concept of this method has already been described, leading to a qualitative monitoring of transport. We now work at defining a reconstitution procedure amenable to now quantify the rate of transport. To do so we take extreme care to precisely determine the efficiency of protein reconstitution and the type of lipids component used to perform liposomes. I will present the roadmap towards the rational, step-by-step, reconstitution of the MexA-MexB-OprM efflux pump as well as the methodologies that are undertaken to measure the velocity of transport, and possible perspectives regarding the screening of efflux pump inhibitors.
From the 77 posters that were presented on-site, 3 were selected as the winners after a shortlist round by popular vote, followed by a selection round by the conference organisers.
Resolving noise-control conflict by gene duplication
Authors: Michal Chapal, Sefi Mintzer, Sagie Brodsky, Miri Carmi, Naama Barkai, Weizmann Institute of Science, Israel
Gene duplication promotes adaptive evolution in two principle ways: allowing one duplicate to evolve a new function and resolving adaptive conflicts by splitting ancestral functions between the duplicates. In an apparent departure from both scenarios, low-expressing transcription factor duplicates commonly regulate similar sets of genes and act in overlapping conditions. To examine for possible benefits of such apparently redundant duplicates, we examined the budding yeast duplicated stress regulators Msn2 and Msn4. We show that Msn2,4 indeed function as one unit, inducing the same set of target genes in overlapping conditions, yet this two-factor composition allows its expression to be both environmental-responsive and with low-noise, thereby resolving an adaptive conflict that inherently limits expression of single genes. Our study exemplified a new model for evolution by gene duplication whereby duplicates provide adaptive benefit through cooperation, rather than functional divergence: attaining two-factor dynamics with beneficial properties that cannot be achieved by a single gene.
Deep learning on single-cell ATAC-seq data to decipher enhancer logic
Authors: Ibrahim Ihsan Taskiran, Liesbeth Minnoye, Carmen Bravo Gonzalez-Blas, Sara Aibar Santos, Gert Hulselmans, Valerie Christiaens, Stein Aerts KU Leuven – VIB, Belgium
Single-cell ATAC-seq provides new opportunities to study gene regulation in heterogeneous cell populations such as complex tissues or dynamic processes. We recently developed a probabilistic topic modeling approach, called cisTopic, to predict regulatory topics and sets of co-accessible enhancers from scATAC-seq data. Here, we apply deep learning approaches to analyze these sets of co-accessible enhancers, with the goal to predict the spatiotemporal pattern of enhancer accessibility directly from the enhancer sequence. We trained different types of Artificial Neural Networks, including a hybrid model that combines Convolutional and Recurrent Neural Networks. By applying this approach to a cohort of melanoma patient samples and Drosophila eye disc, we show that key transcription factors can be identified from the convolutional filters. In addition, we use the trained model to analyze the motif architecture in enhancers, such as motif combinations and relationship to nucleosome preferences. We furthermore exploit network explaining methods to predict the impact of somatic mutations, using publicly available SNP databases and in-house whole genome sequencing of inbred fly lines. Currently, to validate our models we are testing (mutated) synthetic cell state specific enhancers using massively parallel enhancer reporter assays (MPRA). In conclusion, training deep learning models on single-cell epigenomics data sets has multiple applications to understand the underlying enhancer logic and decipher gene expression programs.
ROADdt: Regulation network remodeling along disease development trajectories
Authors: Celine Sin, Jörg Menche CeMM Research Center for Molecular Medicine of the Austrian Academy of Sciences, Austria
The human body is comprised of over 200 different cell types varying in size, shape, and function. The differentiation and subsequent maintenance of these different phenotypic states are governed by complex gene regulatory networks that dynamically orchestrate the activation and deactivation of genes. Abnormalities in these networks may lead to dysfunctional expression programs, e.g. uncontrolled cell proliferation. In order to understand the conditions resulting in disease, we must understand the underlying gene regulatory networks governing
the gene expression program. As cells move through the differentiation space, the networks that govern gene regulation are remodeled in order to achieve the appropriate gene expression program. While statistical physics and network theory have demonstrated numerous relationships between the structure of networks and the dynamic processes that act on them,
few studies link these mathematically rigorous principles to gene regulatory networks, none at the level of cell-trajectory-states. The overall goal of this project is to understand the fundamental architecture of gene regulatory networks associated with cell differentiation processes in disease. We hypothesize that the gene regulatory networks of different
cell-trajectory-states along the differentiation trajectory – e.g. transitory, branching, or terminal states – are each characterized by distinct structural features. I will present our first steps in this direction, starting from single-cell RNA seq profiles of tumors. Ultimately, we expect that detailed characterization of the gene regulatory networks in these disease processes will reveal basic principles applicable to other diseases and cell developmental processes.