Richa Tiwari

A supervised machine learning approach of extracting coexpression relationship among genes from literature

It is vital to develop automatic information extraction systems to help researchers cope up with the vast amount of data available on the Internet. In this paper, we describe a framework to extract precise information about coexpression relationship among genes, from published literature using a supervised machine learning approach. We use a graphical model, Dynamic Conditional Random Fields (DCRFs), for training our classifier. Our approach is based on semantic analysis of text to classify the predicates describing coexpression relationship rather than detecting the presence of keywords. We compared our results of sentence classification with the baseline technique of word matching and a Naïve Bayes classification algorithm. Our framework outperformed the baseline by almost 45%, with DCRFs showing superior performance to Naïve Bayes.

Comparison of microarray image analysis software

Microarray is very useful and vastly growing technology that helps in analyzing genetic data. Because of the advancement in this field there is lot of genetic data being produced everyday and these data are needed to be interpreted and analyzed properly to get accurate results. With the advancement in computer technologies and software, we now analyze microarray slides by computers. There are many open sources as well as commercially available software that help us to analyze microarray images. Microarray image analysis can be divided into three stages-Addressing or Gridding of spots, Segmentation and finally Background and Foreground Intensity Extraction. This paper tries to compare few of the available microarray image analysis software on the basis of the above mentioned techniques used and also on the efficiency and ease of use of these software.

Identifying features to improve real time clustering and domain blacklisting

Feature analysis is an important task in the area of information extraction. Appropriate features give improved performance for any classification or clustering algorithm. In this paper we try to analyze different features that can be used to cluster spam emails at real time and thus improve IP blacklisting. Domain blacklisting becomes easy when these features are used because masses of IP address get grouped easily. We have explored several features in this paper like sender and subject of the email; email attachments, stylistic and semantic features. These features ensure appropriate clustering of spam originating from dominant hosts. We compute the effectiveness of these features in terms of how well they group emails, gather domain/IP information and thus improve domain blacklisting.

A Supervised Machine Learning Approach of Extracting and Ranking Published Papers Describing Coexpression Relationships among Genes

In this chapter, we describe a framework to extract information about coexpression relationships among genes from published literature using a supervised machine learning approach, and later rank those papers to provide users with a complete specialized information retrieval system. We use Dynamic Conditional Random Fields (DCRFs), for training our classification model. Our approach is based on semantic analysis of text to classify the predicates describing coexpression rather than detecting the presence of keywords. Our framework outperformed the baseline by almost 52%, with DCRFs showing superior performance to Bayes Net, SVM, and Naive Bayes classification algorithm. In our second experiment, the comparison of our ranked results to that of PubMed and Google demonstrates that our proposed model performs better than both in distinguishing a positive paper from a negative paper. In conclusion, this chapter describes a specialized classification and ranking framework that can retrieve articles that discuss coexpression among genes.

A Data Mining Method to Extract and Rank Papers Describing Coexpression Predicates Semantically

Information management and extraction in the field of biomedical research has become a requirement with the rapid increase in the amount of data being published in this area. In this paper, a graphical model, Conditional Random Fields has been used to extract a particular gene-gene relationship called “coexpression” from the existing literature. First, a Conditional Random Fields based model has been trained and tested on full-length papers downloaded from PubMed, to label the predicates that talk about coexpression of genes. Proper local and contextual text features at both word and sentence levels are proposed and extracted during the pre-processing step. The classification performance of the model trained based on the proposed features has been compared with the that of Support Vector Machines, Nearest Neighbor with generalization, and Neural Networks algorithms, and seen to outperform them all. In our second experiment, the proposed ranking scheme, which is based on classification results, is applied to the ranked lists of papers returned by PubMed and Google, respectively. The comparison of our ranked results to that of PubMed and Google demonstrates that our proposed ranking scheme performs better than both in distinguishing a positive paper from a negative paper. In conclusion, this paper describes a specialized classification and ranking framework that can retrieve papers that really talk about coexpression between and among genes based on mining of semantics and not just lexical search.

Extraction of coexpression relationship among genes from biomedical text using dynamic conditional random fields

Text mining tools and algorithms are being successfully used for information extraction especially on large corpus like biomedical publications. These tools not only aid in information extraction but also in forming new theories and relationships between various fields of biomedical research. Extraction of gene-gene or gene-disease relationship is one such application. In this paper, we introduce a method to detect coexpressed genes from text, using the grammatical dependencies among the words within sentences and Dynamic Conditional Random Fields (DCRFs). Determining the coexpression relationship between and among genes can help in identifying important concepts such as the functionality of gene(s) involved, their pathogenic mechanism, and in deciphering protein-protein interactions. This work attempts to extract relevant sentences by labeling the genes involved as well as the word representing the relationship, from full-length papers collected from PubMed. The results obtained were compared with that of Support Vector Machine (SVM) and Nearest Neighbor with generalization (NNge), and have been found to outperform both.