Data infrastructures for AI in medical imaging: a report on the experiences of five EU projects.

Artificial intelligence (AI) is transforming the field of medical imaging and has the potential to bring medicine from the era of 'sick-care' to the era of healthcare and prevention. The development of AI requires access to large, complete, and harmonized real-world datasets, representative of the population, and disease diversity. However, to date, efforts are fragmented, based on single-institution, size-limited, and annotation-limited datasets. Available public datasets (e.g., The Cancer Imaging Archive, TCIA, USA) are limited in scope, making model generalizability really difficult. In this direction, five European Union projects are currently working on the development of big data infrastructures that will enable European, ethically and General Data Protection Regulation-compliant, quality-controlled, cancer-related, medical imaging platforms, in which both large-scale data and AI algorithms will coexist. The vision is to create sustainable AI cloud-based platforms for the development, implementation, verification, and validation of trustable, usable, and reliable AI models for addressing specific unmet needs regarding cancer care provision. In this paper, we present an overview of the development efforts highlighting challenges and approaches selected providing valuable feedback to future attempts in the area.Key points• Artificial intelligence models for health imaging require access to large amounts of harmonized imaging data and metadata.• Main infrastructures adopted either collect centrally anonymized data or enable access to pseudonymized distributed data.• Developing a common data model for storing all relevant information is a challenge.• Trust of data providers in data sharing initiatives is essential.• An online European Union meta-tool-repository is a necessity minimizing effort duplication for the various projects in the area.

Detection of early seeding of Richter transformation in chronic lymphocytic leukemia.

Richter transformation (RT) is a paradigmatic evolution of chronic lymphocytic leukemia (CLL) into a very aggressive large B cell lymphoma conferring a dismal prognosis. The mechanisms driving RT remain largely unknown. We characterized the whole genome, epigenome and transcriptome, combined with single-cell DNA/RNA-sequencing analyses and functional experiments, of 19 cases of CLL developing RT. Studying 54 longitudinal samples covering up to 19 years of disease course, we uncovered minute subclones carrying genomic, immunogenetic and transcriptomic features of RT cells already at CLL diagnosis, which were dormant for up to 19 years before transformation. We also identified new driver alterations, discovered a new mutational signature (SBS-RT), recognized an oxidative phosphorylation (OXPHOS)high-B cell receptor (BCR)low-signaling transcriptional axis in RT and showed that OXPHOS inhibition reduces the proliferation of RT cells. These findings demonstrate the early seeding of subclones driving advanced stages of cancer evolution and uncover potential therapeutic targets for RT.

The Quest for Orthologs benchmark service and consensus calls in 2020.

The identification of orthologs-genes in different species which descended from the same gene in their last common ancestor-is a prerequisite for many analyses in comparative genomics and molecular evolution. Numerous algorithms and resources have been conceived to address this problem, but benchmarking and interpreting them is fraught with difficulties (need to compare them on a common input dataset, absence of ground truth, computational cost of calling orthologs). To address this, the Quest for Orthologs consortium maintains a reference set of proteomes and provides a web server for continuous orthology benchmarking (http://orthology.benchmarkservice.org). Furthermore, consensus ortholog calls derived from public benchmark submissions are provided on the Alliance of Genome Resources website, the joint portal of NIH-funded model organism databases.

4D Genome Rewiring during Oncogene-Induced and Replicative Senescence.

To understand the role of the extensive senescence-associated 3D genome reorganization, we generated genome-wide chromatin interaction maps, epigenome, replication-timing, whole-genome bisulfite sequencing, and gene expression profiles from cells entering replicative senescence (RS) or upon oncogene-induced senescence (OIS). We identify senescence-associated heterochromatin domains (SAHDs). Differential intra- versus inter-SAHD interactions lead to the formation of senescence-associated heterochromatin foci (SAHFs) in OIS but not in RS. This OIS-specific configuration brings active genes located in genomic regions adjacent to SAHDs in close spatial proximity and favors their expression. We also identify DNMT1 as a factor that induces SAHFs by promoting HMGA2 expression. Upon DNMT1 depletion, OIS cells transition to a 3D genome conformation akin to that of cells in replicative senescence. These data show how multi-omics and imaging can identify critical features of RS and OIS and discover determinants of acute senescence and SAHF formation.

BioExcel Building Blocks, a software library for interoperable biomolecular simulation workflows.

In the recent years, the improvement of software and hardware performance has made biomolecular simulations a mature tool for the study of biological processes. Simulation length and the size and complexity of the analyzed systems make simulations both complementary and compatible with other bioinformatics disciplines. However, the characteristics of the software packages used for simulation have prevented the adoption of the technologies accepted in other bioinformatics fields like automated deployment systems, workflow orchestration, or the use of software containers. We present here a comprehensive exercise to bring biomolecular simulations to the "bioinformatics way of working". The exercise has led to the development of the BioExcel Building Blocks (BioBB) library. BioBB's are built as Python wrappers to provide an interoperable architecture. BioBB's have been integrated in a chain of usual software management tools to generate data ontologies, documentation, installation packages, software containers and ways of integration with workflow managers, that make them usable in most computational environments.

Four simple recommendations to encourage best practices in research software.

Scientific research relies on computer software, yet software is not always developed following practices that ensure its quality and sustainability. This manuscript does not aim to propose new software development best practices, but rather to provide simple recommendations that encourage the adoption of existing best practices. Software development best practices promote better quality software, and better quality software improves the reproducibility and reusability of research. These recommendations are designed around Open Source values, and provide practical suggestions that contribute to making research software and its source code more discoverable, reusable and transparent. This manuscript is aimed at developers, but also at organisations, projects, journals and funders that can increase the quality and sustainability of research software by encouraging the adoption of these recommendations.

PMut: a web-based tool for the annotation of pathological variants on proteins, 2017 update.

We present here a full update of the PMut predictor, active since 2005 and with a large acceptance in the field of predicting Mendelian pathological mutations. PMut internal engine has been renewed, and converted into a fully featured standalone training and prediction engine that not only powers PMut web portal, but that can generate custom predictors with alternative training sets or validation schemas. PMut Web portal allows the user to perform pathology predictions, to access a complete repository of pre-calculated predictions, and to generate and validate new predictors. The default predictor performs with good quality scores (MCC values of 0.61 on 10-fold cross validation, and 0.42 on a blind test with SwissVar 2016 mutations). The PMut portal is freely accessible at http://mmb.irbbarcelona.org/PMut. A complete help and tutorial is available at http://mmb.irbbarcelona.org/PMut/help.

BIGNASim: a NoSQL database structure and analysis portal for nucleic acids simulation data.

Molecular dynamics simulation (MD) is, just behind genomics, the bioinformatics tool that generates the largest amounts of data, and that is using the largest amount of CPU time in supercomputing centres. MD trajectories are obtained after months of calculations, analysed in situ, and in practice forgotten. Several projects to generate stable trajectory databases have been developed for proteins, but no equivalence exists in the nucleic acids world. We present here a novel database system to store MD trajectories and analyses of nucleic acids. The initial data set available consists mainly of the benchmark of the new molecular dynamics force-field, parmBSC1. It contains 156 simulations, with over 120 µs of total simulation time. A deposition protocol is available to accept the submission of new trajectory data. The database is based on the combination of two NoSQL engines, Cassandra for storing trajectories and MongoDB to store analysis results and simulation metadata. The analyses available include backbone geometries, helical analysis, NMR observables and a variety of mechanical analyses. Individual trajectories and combined meta-trajectories can be downloaded from the portal. The system is accessible through http://mmb.irbbarcelona.org/BIGNASim/. Supplementary Material is also available on-line at http://mmb.irbbarcelona.org/BIGNASim/SuppMaterial/.

BioSWR--semantic web services registry for bioinformatics.

Despite of the variety of available Web services registries specially aimed at Life Sciences, their scope is usually restricted to a limited set of well-defined types of services. While dedicated registries are generally tied to a particular format, general-purpose ones are more adherent to standards and usually rely on Web Service Definition Language (WSDL). Although WSDL is quite flexible to support common Web services types, its lack of semantic expressiveness led to various initiatives to describe Web services via ontology languages. Nevertheless, WSDL 2.0 descriptions gained a standard representation based on Web Ontology Language (OWL). BioSWR is a novel Web services registry that provides standard Resource Description Framework (RDF) based Web services descriptions along with the traditional WSDL based ones. The registry provides Web-based interface for Web services registration, querying and annotation, and is also accessible programmatically via Representational State Transfer (REST) API or using a SPARQL Protocol and RDF Query Language. BioSWR server is located at http://inb.bsc.es/BioSWR/and its code is available at https://sourceforge.net/projects/bioswr/under the LGPL license.

Cystinuria-specific rBAT(R365W) mutation reveals two translocation pathways in the amino acid transporter rBAT-b0,+AT.

Apical reabsorption of dibasic amino acids and cystine in kidney is mediated by the heteromeric amino acid antiporter rBAT/b(0,+)AT (system b(0,+)). Mutations in rBAT cause cystinuria type A, whereas mutations in b(0,+)AT cause cystinuria type B. b(0,+)AT is the catalytic subunit, whereas it is believed that rBAT helps the routing of the rBAT/b(0,+)AT heterodimeric complex to the plasma membrane. In the present study, we have functionally characterized the cystinuria-specific R365W (Arg(365)-->Trp) mutation of human rBAT, which in addition to a trafficking defect, alters functional properties of the b(0,+) transporter. In oocytes, where human rBAT interacts with the endogenous b(0,+)AT subunit to form an active transporter, the rBAT(R365W) mutation caused a defect of arginine efflux without altering arginine influx or apparent affinities for intracellular or extracellular arginine. Transport of lysine or leucine remained unaffected. In HeLa cells, functional expression of rBAT(R365W)/b(0,+)AT was observed only at the permissive temperature of 33 degrees C. Under these conditions, the mutated transporter showed 50% reduction of arginine influx and a similar decreased accumulation of dibasic amino acids. Efflux of arginine through the rBAT(R365W)/b(0,+)AT holotransporter was completely abolished. This supports a two-translocation-pathway model for antiporter b(0,+), in which the efflux pathway in the rBAT(R365W)/b(0,+)AT holotransporter is defective for arginine translocation or dissociation. This is the first direct evidence that mutations in rBAT may modify transport properties of system b(0,+).