BioTextRetriever: A Tool to Retrieve Relevant Papers

An Architecture For An Information Retrieval System

The overall goal of our work is to implement a web based search tool that receives a set of genomic or proteomic sequences and returns an ordered set of papers relevant to the study of such sequences. The initial set of sequences is supplied by a biologist together with a set of relevant keywords and an e-value (an e-value is a statistic to estimate the significance of a match between 2 sequences). These three items are the input for BioTextRetriever (Gonçalves, Camacho, & Oliveira, 2011) as can be seen in Figure 1. Figure 1 presents a summary of our approach that we will now describe in detail. In the following description we use NCBI as the sequence Data Base.

Figure 1.

Sequence of steps executed by BioTextRetriever when the user provides a set of initial DNA/protein sequences

In Step 1, the user (a biologist researcher) provides an initial of sequences, optionally a list of keywords, and an e-value. With these three items (sequences, keywords and e-value) and using the NCBI BLAST tool we collect a set of similar sequences together with the paper references associated to them. We could also use Ensembl with the same inputs because Ensembl may return a different set of papers references. However for the proposed work we have only used the NCBI database.