DyNet LS

DyNet LS intelligence extraction is made available through DyNet's semantic module that recognises terms and concepts such as: genes, proteins, diseases and rare diseases, drugs, chemicals, biological processes, cellular components, medicinal products. Moreover DyNet's semantics can be customised to add vocabularies or thesauri to search in-house or commercially available knowledge bases.

The user front end is intuitive and based on common commercial features: left click/selection, right click/options.

DyNet LS will display a network of related concepts that represent co-occurrences in science data. Every node in the graph will have different size and connections based on its frequency within the analysed data. Users can analyse the graph by applying topological measurements such as centrality, weighted centrality, core, betweenness, clustering coefficient, relative intensity, etc. These parameters are very useful in validating a given node/vertex in terms of importance, connections to other nodes/vertices, degree of embeddedness, extent to which the node/vertex exerts its importance by spreading knowledge/data/importance (please review the technical manual for a detailed and complete explanation of the aforementioned features).

DyNet has been successfully applied to publications, patents, clinical trials, RSS feeds in the pharma sector and peer selection, other biological databases like KEGG, Microarray databases, outputs from HTS experiments on multidimensional level.

Default databases in DyNet LS are: pubmed, patents, clinical trials

Try it!

For a more comprehensive understanding of DyNet LS, ATA offers a 15 day FREE trial. To get a free trial download the software here.

Why DyNet LS

DyNet LS has been developed to facilitate analysis of biodata such as scientific publications, patents, clinical trials, biological & chemical databases and other relational knowledge bases. DyNet LS and its semantic tool will extract key concepts such as genes, proteins, drugs, chemicals, diseases and rare diseases, molecular functions, biological processes, cellular components, medicinal products and other adopted thesauri while displaying their co-occurrences in order to unveil relations such as genotype-phenotype, gene-disease, protein-chemicals-drugs and many others.

What can you do?

• Perform pubmed searches on authors (result=a network of relations among scientists with MESH terms).
• Analyse scientific communities, their reach and collaborations
• Perform topological analysis to understand performances in terms of publications, last authorships, outcomes and degree of embeddedness and centrality of a given scientist, institution or related biomedical concepts
• Extract from pubmed Concepts, Co-occurences and their relations and display them as a network
• Unveil protein-protein interactions and signalling
• Unveil gene-protein feedbacks
• Perform semantic analysis of publications for the identifications of drug targets, key proteins and scaffolds, signalling events, association gene-disease, gene-drug, protein-disease, protein-drug, etc.
• Perform technology landscape analysis within DyNet's patent database (US-EP-WO, 1970-2007)
• Analyse bioconcepts and technology for a given company/organisation on EU-US-WO patents in a dynamic fashion (over time) or compare two companies in terms of their drugs, targets, bioconcepts, medicinal products, diseases
• Compare two networks
• Perform patent citation analysis
• Search ClinicalTrials.gov database with DyNet's semantic tool and analyse who is working on what at a glance
• Assess drug target scientific validation
• Perform peer selection for funding applications
• Select potential collaborators
• Import data in text and XML formats
• Export data and attributes in text format or PNG, VRML, SVG
• Perform analysis of DyNet's RSS pharma feeds database to scout for partnerships, deals, regulatory affairs and check for relations among diseases, durgs, targets, medicinal products and other bioconcepts
• Search DATABiotech, ATA's proprietary global database dedicated to the life sciences sector with more than 27,000 entries including: pharma, biotech, biotech suppliers, Public research organisations, VCs and many more. DATABiotech reports on contact and profile info, deals, technology and products, corporate alliances, news highlights, patents, publications, key scientists, MESH terms.
• Customise your DyNet license with in-house proprietary or commercial databases

Semantics

ATA has implemented a semantic module that extracts intelligence from scientific literature, patents, XML, text data, abstracts, RSS feeds or any proprietary or in-house database. The module has been also designed to accommodate customised vocabularies and thesauri. Concepts that co-occur in scientific literature, patents, biodata, medical abstracts clinical trials are displayed graphically to analyse its scientific and technological landscape including KOL (key opinion leaders) and KOO (Key opinion operators)

How does it work?

DyNet's semantic approach adopts a cascaded approach to Name Entity Recognition (NER) and markup in scientific data. DyNet keeps the annotations made for example by publishers (ie: title, author, section, headings, citations, references, etc) and inserts XML annotations during text analysis. XML annotations is performed by looking for terms in referential databases like GO, SWISSPROT, CHEBI, DRUGBANK.

DyNet's approach to NER can be divided in three steps:

1. Term recognition → CONCEPT
2. Term categorisation → GENE, PROTEIN, SPECIES, etc
3. Mapping → LINK TO REFRENTIAL DATABASES

Users can decide the order in which concepts are identified and prioritise against other concepts occurring in their reference database. That is users, after markup of text, term identification, disambiguation, can decide to order term identification with GENES preceding DRUG terms, PROTEINS, CHEMICALS or any other desired order.

Available Thesauri

GENES (GO), PROTEINS (Uniprot-Swissprot), DRUGS (drugbank); DISEASES (Healthcentral), SPECIES (NCBI), CHEMICALS (Chebi), BIOLOGICAL PROCESSES (QuickGO), CELLULAR COMPONENTS (QuickGO), OMIM (NCBI), MOLECULAR FUNCTION (QuickGO), MEDICINAL PRODUCTS (FDA, EMEA), GENE NAMES (HUGO).

Disambiguation, homonyms, acronyms are dealt using other meta-thesauri especially for terms that are also common english terms.