Introduction to the basic SCAIView functions

Most of the current biomedical knowledge from these large text collections is present in unstructured scientific text (journal publications, text fields in databases, e.g. more than 20 Mio. Documents in PubMed).

SCAIView provides users with full text and biomedical concept search capabilities, which are supported by large biomedical terminologies and ontologies, processed together with outstanding text mining technologies. Using machine learning and dictionary-based Named Entity Recognition (NER), SCAIView identifies information about genes, drugs, SNPs and other Life Science entities in MEDLINE abstracts and extracts this information. SCAIView uses a multi-threaded Lucene Index to allow semantic and ontological search on unstructured (text) data. Complex queries such as “what drugs are mentioned in the context of Alzheimers disease”? or “what genes are co-mentioned with Diabetes and are on the insulin signalling pathway”? can be asked in a user-friendly, intuitive way.

Please follow the following pages of the "First Steps" to get in touch with SCAIView!

Most of the current biomedical knowledge from these large text collections is present in unstructured scientific text (journal publications, text fields in databases, e.g. more than 20 Mio. Documents in PubMed).
SCAIView provides users with full text and biomedical concept search capabilities, which are supported by large biomedical terminologies and ontologies, processed together with outstanding text mining technologies. Using machine learning and dictionary-based Named Entity Recognition (NER), SCAIView identifies information about genes, drugs, SNPs and other Life Science entities in MEDLINE abstracts and extracts this information. SCAIView uses a multi-threaded Lucene Index to allow semantic and ontological search on unstructured (text) data. Complex queries such as “what drugs are mentioned in the context of Alzheimers disease”? or “what genes are co-mentioned with Diabetes and are on the insulin signalling pathway”? can be asked in a user-friendly, intuitive way.

The new SCAIView entry page

Version 0.9.5 has now a new look and improved functionality, and is ready to be used by users without any bioinformatics experts background. All you need is some basic experience with Google. In particular, query building is now supported through auto-completion of typed words which results in rapid building of complex queries and fast retrieval of relevant information. The new Graphical User Interface (GUI) offers several new features (see Graphic 1) including the text input field 'Query Builder' and the selection menu 'Show results in'.

First Query

In the Query Builder you type in your search terms. You can use any term you want with a 'full text' search in the knowledge space. This feature behaves like a normal Google search function and results in a list of all occurrences of that search term in the literature. In addition, SCAIView 0.9.5 provides an auto-completion function to suggest a list of classes.

Auto-completion in the Query Builder

The user is free to select the free-text search (select first entry in the column) or to select any class provided by the autocompletion window. The auto-completion function includes all synonyms of the search dictionaries and suggests the classes to be selected. The class search is more intelligent in a sense that it uses domain specific terms and synonyms, and sub-classes thereby generating more specific results. Typing a disease-specific term like ‘Alzh…’ activates the auto-completion functionality which suggests all the available classes containing your typed words.

Further Queries

After editing a search term into the ‘Query Builder’ input field, SCAIView automatically gives you an additional empty input field, to edit further query terms. You can select, in which logical (Boolean) concatenation this new search term has to be evaluated. The standard search operation is a Boolean AND connections shown on the left hand side of the search box. The combination can be changed to OR and NOT through the drop down menu.
You can add or delete terms to your query. To delete a search term the minus button on the right hand side can be used.

SCAIView complex Query building

Query visualization

A new feature enables users to visualize the query or copy/paste queries. To use this feature you must choose the “Show/Hide Query” button to get the ‘Your Query’ input field.

SCAIView Query visualisation

Selection of classes from Search tree

All annotated classes are visualized in the tree on the left hand side of the interface. In the entry page the tree is hidden and can be visualized by clicking the 'Show/Hide Overall Query' button. The tree can be used to browse within the different classes. If a class is selected it will be added to the Query Builder as additional search class. The Class tree offers a search field, which helps to find concepts in the tree structure.

SCAIView Show tree feature

Show Results in

The selection menu 'Show results in' gives the possibility to select the search results in direct document visualization (show documents or overview pages about specific entities such as genes, drugs or diseases. Depending on your setting in the ‘Show Results in’ selection menu, the results of your query will be shown in the “Document view” or “Entity view”. SCAIView changes automatically to the view / tab, accordingly to your selection. You can go back to the “Query tab” to modify your query.

If you choose ‘Documents’ in the ‘Show Results in’ selection menu, results of your query will result in a list of relevant documents (Figure 7). SCAIView changes automatically to the tab ‘Documents’ view. In this tab you have several information and options.:

  • You can switch back to another flag
  • Your Query is shown
  • You can select another view in Show Results
  • ’Select/deselect all entity classes’ or activate specific entity classes to highlight the entities in your relevant documents.
  • The number of found relevant documents accordingly your query. I you have selected an entry in the entity view the selected entity and the corresponding highlighting is shown
  • You can scroll (pagewise) through the list of relevant documents in arbitrary direction using the scroll buttons ‘Next’ and ‘Last’ or select a page of interest.

SCAIView Document View with highlighted entity classes

If you choose a specific dictionary or ontology in the ‘Show Results in’ selection menu, e.g. ‘Human Genes / Proteins’, your query will result in a list of relevant entities (Figure 8). SCAIView changes automatically to the tab ‘Entity’ view.
In this tab you have several information and options.:

  • You can switch back to another flag
  • Your Query is shown
  • You can select another view in Show Results
  • show only those entities which where selected in your query; e.g. if you select genes from a certain pathway, only those genes are shown.
    If this box is not activated all genes of the pathway plus all cooccuring entities are shown.
  • Select table column: select your entity specific configuration of the entity view
  • The number of all relevant entities.
  • A list which is shown pagewise of all these entities relevant for your query in arbitrary direction using the scroll buttons ‘Next’ and ‘Last’ or select a page of interest. Mouse over the entity shows the corresponding definition. If you select one of these entities you switch to the document view and the document with respect to the selected entity and your query are shown. Selecting the opens the analysis view for this entity in the selected query.
  • A list which is shown pagewise of all these entities relevant for your query in arbitrary direction using the scroll buttons ‘Next’ and ‘Last’ or select a page of interest.
    Mouse over the entity shows the corresponding definition
    If you select one of these entities you switch to the document view and the document with respect to the selected entity and your query are shown
    Selecting the opens the analysis view for this entity in the selected query.
  • For each entity the relevance ranking based on relative entropy is given be default.
  • Reference document count states the number of abstracts
  • Doc count states the number of abstracts corresponding to your query where the entity is found

SCAIView Entities View with highlighted entity classes

Analysis View: To see the frequencies of all entities over all found abstracts, return to the Entity tab and click on the analysis icon. You will be directed to the Analysis tab where an overview of the entities and their occurrence frequencies in the abstracts is given.

Analyses View with frequencies of entities for all classes