Rodan Development (Poland)

Abstract Background: Drug target discovery is a vital step in drug development. Novel computational methods taking advantage of multimodal data sets and diverse algorithmic approaches are beginning to provide key insights to inform this process. Comprehensive biomedical knowledge graphs provide a particularly promising paradigm for target discovery and prioritization, being able to encode multi-modal relational evidence without the need for feature engineering or selection. Within this context, target prediction can be formulated as a link prediction problem using knowledge graph embeddings (KGEs), which learn vector representations of graph entities and provide a flexible framework to simultaneously prioritize drug targets and interpret them within the broader context of their underlying biology. Approach: We developed a comprehensive data-driven biomedical knowledge graph spanning over 4 billion relations across over 1 million entities including genes, diseases, drugs, pathways, tissues, cell types and other biomedical entities. Using a subset of this knowledge graph derived directly from molecular networks as well as analyses of large-scale cancer genomics data (including gene and protein expression, CRISPR sensitivity, somatic mutations and hypermethylation), we trained, evaluated and optimized 17 state-of-the-art KGE models, further incorporating a specialized negative sampling strategy and custom stopping criteria. The KGE models were used to predict disease-target associations as well as identify target classes along with their inter- and intra-class interaction profile signatures. Results: Cancer target predictions were evaluated using a benchmark set of known drug targets. The best KGE methods sharing molecular and drug target information across cancer types achieved an average AUPRC of 0.794 and significantly outperformed standard machine learning approaches trained on each cancer separately. KGE gene embeddings formed distinct clusters aligned with target-disease associations and gene-gene interaction profiles. Analysis of these clusters revealed sets of cancer drug targets enriched in distinct biological functions, including immune checkpoints, GPCRs, cell cycle regulators, and ribosomally-associated genes. Many of the predicted new targets were embedded close to genes with established target roles - providing alternative target candidates which share common molecular pathways and/or biological mechanisms with already established targets. Notably, we also identified gene clusters with few or no known targets whose cancer relationships were nevertheless strongly supported by the underlying data, highlighting the potential of our approach to discover target candidates within new functional classes. Citation Format: Bob Zimmermann, Radosław Bielecki, Roy Ronen, Janusz Dutkowski. Knowledge graph AI-based prioritization of drug target candidates across 503 cancers [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_2):Abstract nr LB112.

O texto discute a relação entre subsídios e proteção comercial nos mercados globais de produtos agropecuários, destacando a importância histórica das rodadas do Acordo Geral sobre Tarifas e Comércio (GATT) na regulação dessas práticas. Embora a Rodada Uruguai tenha representado um avanço na regulamentação de subsídios à produção e exportação de produtos agropecuários, as expectativas para a Rodada de Doha não se concretizaram, resultando em um enfraquecimento do multilateralismo e um aumento do foco em acordos bilaterais entre grandes atores como a União Europeia (UE), Estados Unidos e China. O texto tem como objetivo analisar a trajetória dos subsídios agropecuários da UE no período de 2000 a 2020 e avaliar a proteção comercial exercida por este bloco em relação às importações de alimentos, fornecendo insights relevantes para formuladores de políticas públicas e privadas, especialmente do ponto de vista brasileiro. O trabalho é estruturado em cinco seções, incluindo uma introdução, uma seção sobre a importância da UE no contexto global agropecuário, metodologia, resultados e considerações finais.

The task of an information integration system is to combine data residing at different sources, providing the user with a unified view of them, called global schema. Users formulate queries over the global schema, and the system suitably queries the sources, providing an answer to the user, who is not obliged to have any information about the sources. Recent developments in IT such as the expansion of the Internet and the World Wide Web, have made available to users a huge number of information sources, generally autonomous, heterogeneous and widely distributed: as a consequence, information integration has emerged as a crucial issue in many application domains, e.g., distributed databases, cooperative information systems, data warehousing, or on-demand computing. Recent estimates view information integration to be a $10 Billion market by 2006 [14].

Many requirements for a business process depend on the workflow execution data that includes common data for all the population of processes, state of resources, state of processes, etc. The natural way to specify and implement such requirements is to put them into the process definition. In order to do it, we need: (1) a generalised workflow metamodel that includes data on the workflow environment, process definitions, and process execution; (2) a powerful and flexible query language addressing the metamodel; (3) integration of a query language with a business process definition language. In this paper the mentioned workflow metamodel together with the business process query language BPQL is presented. BPQL is integrated with the XML Process Definition Language (XPDL) increasing significantly its expressiveness and flexibility. We also present practical results for application of the proposed language in the OfficeObjects® WorkFlow system.