Hugo Costa

Development of Text Mining Tools for Information Retrieval from Patents

Biomedical literature is composed of an ever increasing number of publications in natural language. Patents are a relevant fraction of those, being important sources of information due to all the curated data from the granting process. However, their unstructured data turns the search of information a challenging task. To surpass that, Biomedical text mining (BioTM) creates methodologies to search and structure that data. Several BioTM techniques can be applied to patents. From those, Information Retrieval is the process where relevant data is obtained from collections of documents. In this work, a patent pipeline was developed and integrated into @Note2, an open-source computational framework for BioTM. This integration allows to run further BioTM tools over the patent documents, including Information Extraction processes as Named Entity Recognition or Relation Extraction.

Development of a Machine Learning Framework for Biomedical Text Mining

Biomedical text mining (BTM) aims to create methods for searching and structuring knowledge extracted from biomedical literature. Named entity recognition (NER), a BTM task, seeks to identify mentions to biological entities in texts. Dictionaries, regular expressions, natural language processing and machine learning (ML) algorithms are used in this task. Over the last years, @Note2, an open-source software framework, which includes user-friendly interfaces for important tasks in BTM, has been developed, but it did not include ML-based methods. In this work, the development of a framework, BioTML, including a number of ML-based approaches for NER is proposed, to fill the gap between @Note2 and state-of-the-art ML approaches. BioTML was integrated in @Note2 as a novel plug-in, where Hidden Markov Models, Conditional Random Fields and Support Vector Machines were implemented to address NER tasks, working with a set of over 60 feature types used to train ML models. The implementation was supported in open-source software, such as MALLET, LibSVM, ClearNLP or OpenNLP. Several manually annotated corpora were used in the validation of BioTML. The results are promising, while there is room for improvement.

Automating the Extraction of Essential Genes from Literature

The construction of repositories with curated information about gene essentiality for organisms of interest in Biotechnology is a very relevant task, mainly in the design of cell factories for the enhanced production of added-value products. However, it requires retrieval and extraction of relevant information from literature, leading to high costs regarding manual curation. Text mining tools implementing methods addressing tasks as information retrieval, named entity recognition and event extraction have been developed to automate and reduce the time required to obtain relevant information from literature in many biomedical fields. However, current tools are not designed or optimized for the purpose of identifying mentions to essential genes in scientific texts.

Development of an information retrieval tool for biomedical patents

BACKGROUND AND OBJECTIVE The volume of biomedical literature has been increasing in the last years. Patent documents have also followed this trend, being important sources of biomedical knowledge, technical details and curated data, which are put together along the granting process. The field of Biomedical text mining (BioTM) has been creating solutions for the problems posed by the unstructured nature of natural language, which makes the search of information a challenging task. Several BioTM techniques can be applied to patents. From those, Information Retrieval (IR) includes processes where relevant data are obtained from collections of documents. In this work, the main goal was to build a patent pipeline addressing IR tasks over patent repositories to make these documents amenable to BioTM tasks. METHODS The pipeline was developed within @Note2, an open-source computational framework for BioTM, adding a number of modules to the core libraries, including patent metadata and full text retrieval, PDF to text conversion and optical character recognition. Also, user interfaces were developed for the main operations materialized in a new @Note2 plug-in. RESULTS The integration of these tools in @Note2 opens opportunities to run BioTM tools over patent texts, including tasks from Information Extraction, such as Named Entity Recognition or Relation Extraction. We demonstrated the pipelines main functions with a case study, using an available benchmark dataset from BioCreative challenges. Also, we show the use of the plug-in with a user query related to the production of vanillin. CONCLUSIONS This work makes available all the relevant content from patents to the scientific community, decreasing drastically the time required for this task, and provides graphical interfaces to ease the use of these tools.

Reconstructing transcriptional Regulatory Networks using data integration and Text Mining

Transcriptional Regulatory Networks (TRNs) are powerful tool for representing several interactions that occur within a cell. Recent studies have provided information to help researchers in the tasks of building and understanding these networks. One of the major sources of information to build TRNs is biomedical literature. However, due to the rapidly increasing number of scientific papers, it is quite difficult to analyse the large amount of papers that have been published about this subject. This fact has heightened the importance of Biomedical Text Mining approaches in this task. Also, owing to the lack of adequate standards, as the number of databases increases, several inconsistencies concerning gene and protein names and identifiers are common. In this work, we developed an integrated approach for the reconstruction of TRNs that retrieve the relevant information from important biological databases and insert it into a unique repository, named KREN. Also, we applied text mining techniques over this integrated repository to build TRNs. However, was necessary to create a dictionary of names and synonyms associated with these entities and also develop an approach that retrieves all the abstracts from the related scientific papers stored on PubMed, in order to create a corpora of data about genes. Furthermore, these tasks were integrated into @Note, a software system that allows to use some methods from the Biomedical Text Mining field, including an algorithms for Named Entity Recognition (NER), extraction of all relevant terms from publication abstracts, extraction relationships between biological entities (genes, proteins and transcription factors). And finally, extended this tool to allow the reconstruction Transcriptional Regulatory Networks through using scientific literature.

Extracting kinetic information from literature with KineticRE.

To better understand the dynamic behavior of metabolic networks in a wide variety of conditions, the field of Systems Biology has increased its interest in the use of kinetic models. The different databases, available these days, do not contain enough data regarding this topic. Given that a significant part of the relevant information for the development of such models is still wide spread in the literature, it becomes essential to develop specific and powerful text mining tools to collect these data. In this context, this work has as main objective the development of a text mining tool to extract, from scientific literature, kinetic parameters, their respective values and their relations with enzymes and metabolites. The approach proposed integrates the development of a novel plug-in over the text mining framework @Note2. In the end, the pipeline developed was validated with a case study on Kluyveromyces lactis, spanning the analysis and results of 20 full text documents.

A Text Mining Approach for the Extraction of Kinetic Information from Literature

Systems biology has fostered interest in the use of kinetic models to better understand the dynamic behavior of metabolic networks in a wide variety of conditions. Unfortunately, in most cases, data available in different databases are not sufficient for the development of such models, since a significant part of the relevant information is still scattered in the literature. Thus, it becomes essential to develop specific and powerful text mining tools towards this aim. In this context, this work has as main objective the development of a text mining tool to extract, from scientific literature, kinetic parameters, their respective values and their relations with enzymes and metabolites. The pipeline proposed integrates the development of a novel plug-in over the text mining tool @Note2. Overall, the results validate the developed approach.