is organized around the idea of projects
, which consist of curation annotation tasks
, performed on a collection of documents
, by a team of one or more curators
, and considering a pre-defined set of concept
is designed to be simple by applying user-friendly interactions. Thus, it is highly focused on the document and respective information annotation, containing three main interaction areas:
- Navigation and action toolbar: provides the ability to navigate through projects and respective documents, and access Egas integrated features;
- Document and annotations viewer: presents the document with in-line concept and relation annotations;
- Annotations visualization filters: enables filtering concepts and relations presented in the document viewer.
Each interaction area contains action components
that provide access to all features integrated in Egas
- Project management: manage and access project configurations, namely users, concepts, relations and annotation statistics;
- Project and document navigators: navigate through different projects and respective documents;
- Processing tools: access integrated processing tools, such as import, export and automatic annotation services;
- Account management: manage account characteristics;
- Document switcher: easily switch between documents in a project;
- Concepts visualization filter: select concept annotations presented in the document viewer;
- Relations visualization filter: select relation annotations presented in the document viewer.
Document's text is split into highlighted sentences
to simplify visualization and promote focused text analysis and improved information extraction processes. Accordingly, Egas
information annotation of concepts
are highlighted with colored boxes
specific for each concept type. Moreover, due to biomedical terminology complexity, nested concept names
are supported and carefully represented through colored boxes within colored boxes.
To simplify annotated concepts analysis, users can select the concepts to visualize
in the document viewer. By unchecking the checkbox of a specific concept, Egas
removes the colored boxes of that concept from the document viewer, "cleaning" the document representation and making the analysis more focused.
are presented below each sentence using directional lines
tagged with the relation type. The colored boxes connected by the relation markup are placed under the concepts that participate in the relation, which are colored with the same color as the respective concept, making it easy to identify the entire relation.
Selective visualization of relations
is also supported, applying a strategy similar to the concepts visualization filtering.
Project navigation is performed through the project navigator
component, which lists public and private projects that are accessible to the user.
In this list, the following information is provided for each project:
- and Highlighted background: project is private;
- : project is public;
- : user is administrator of the project.
Document navigation is performed using two different components: document navigator and switcher. The document navigator
lists the documents available in the project, which allows the user to easily choose to view a specific document of the project.
The document switcher
component allows easy and fast document switching, through two different actions:
- : shows the previous document;
- : shows the next document.
is one of the most important features of Egas
, which provides easy and interactive annotation of concepts
. To add annotations, users must be registered in Egas and make part of the annotation project (below we describe how the users' invitation process works).
Concept names annotation
is performed in three simple steps:
- Select chunk of text;
- Select concept type;
- The concept name annotation is highlighted with the respective concept color.
a previously annotated concept name, right click on the highlighted concept name and choose the option remove.
Concept relations annotation
is performed in four simple steps:
- Keep pressing the "Alt" keyboard key;
- Click on the first concept name;
- Click on the second concept name and select the type of relation;
- The relation is highlighted below the sentence with the respective relation direction.
a previously annotated relation, right click on the highlighted relation and choose the option remove.
Import allows project administrators to add documents to the currently selected project, which can be performed in three different manners:
- Local: add documents stored in your machine;
- Remote: add documents using specific identifiers from remote servers;
- Search: add documents by searching publicly available and indexed literature.
Local import allows users to provide documents stored in their computer in three formats:
- Text: "*.txt" files with raw text documents;
- A1: "*.txt" files with raw text and "*.a1", "*.a2" or "*.ann" files with concept and/or relation annotations. Please visit http://brat.nlplab.org/standoff.html for further information;
- BioC: "*.xml" file with raw text and concept and relation annotations. Please visit http://bioc.sourceforge.net for further information.
After selecting the files, Egas
also allows selecting the documents to import.
Remote import allows users to add documents from the remote services through a list of unique identifiers:
- PubMed: provide PubMed identifiers (PMIDs) to import abstracts of articles;
- PubMed Central: provide PubMed Central identifiers (PMCs) to import full-text articles.
After selecting the desired documents, Egas
presents the list of documents for further confirmation and selection.
Search allows users to import documents by searching remote services that already have publicly available literature indexed. The following remote services are currently supported:
- PubMed: search and import abstracts of articles;
- PubMed Central: search and import full-text articles.
Note that the provided queries are directly provided to remote services, which allows logic operators such as "AND" and "OR".
After submitting the query, Egas
presents a ranked list of documents, and allows selecting the documents that the user want to import.
allows users to store documents and respective annotations in their local machines. Two different formats are supported:
- A1: for each document, provides a "*.txt" file with the text and a "*.ann" file with concept and relation annotations;
- BioC: provides an "*.xml" file with documents' respective text, concept and relation annotations.
allows selecting the documents to be exported. In the end, the generated files are provided in a compressed "ZIP" file.
provide access to solutions that perform automatic annotation of concepts and/or relations
. Currently, the following services are supported:
- Protein-protein interaction: automatic recognition of proteins and respective interactions;
Automatic recognition of protein-protein interactions is performed through this service, providing the following annotations:
- Concept annotations
- Gene or protein: gene or protein concept names;
- Active words: helper words that may indicate the presence of protein-protein interactions in the sentence;
- Relation annotations
- Equivalent relations: relations between annotated protein names that represent the same concept;
- Protein-protein interactions: relations that represent protein-protein interactions between annotated proteins.
allows users to select the documents that they want to annotate using the automatic service.
allows administrators to configure essential project characteristics, such as users, target concept and relation annotations, and access various statistics regarding the annotation process.
allows inviting and removing users from the project. Inviting is performed through an e-mail-based invitation system, which requires the following steps:
- Invite user by providing an e-mail. User may be invited as a curator or administrator;
- Invite is added to the users list with a symbol;
- Provided e-mail already exists in Egas
- Yes: user just needs to accept the invitation;
- No: user must register in Egas and set a password.
- After invitation being accepted, the pending invite is converted to an active user.
At any time, an administrator is able to remove a user from the project or revoke a previous invite.
allows administrators to specify the target concepts of the project:
- Add a new concept specifying its name and highlighting color;
- : remove an existing concept if no annotations of such concept exist in the project;
- : edit an existing concept by changing its name and/or color.
allows administrators to specify the target relations of the project:
- Add a new relation specifying the target concepts and type of interaction;
- : remove an existing relation if no annotations of such relation exist in the project;
- : edit an existing relation by changing the interaction type.
allows administrators to collect detailed information regarding the annotation process, namely:
- Curation time per:
- User and article.
- Annotated concepts per:
- Annotated relations per:
also allows exporting
collected data into Comma Separated Value (CSV) files for further analysis and processing.