e-LiSe--an online tool for finding needles in the '(Medline) haystack'.

Area of Science:

• Bioinformatics
• Computational Biology
• Biomedical Informatics

Background:

• Literature databases contain known and novel biological associations.
• Selecting true correlations from random chance is a major challenge in bio-medical research, especially for low-frequency associations.
• Existing methods struggle with identifying significant, understudied biological relationships.

Purpose of the Study:

• To develop a novel bio-linguistic statistical method for selecting true biological correlations.
• To address the challenge of identifying significant, low-frequency associations in biomedical literature.
• To implement this method in a user-friendly web application.

Main Methods:

• A novel bio-linguistic statistical approach based on Z-score.
• Development of a web-based application named 'e-LiSe'.
• Statistical analysis to distinguish true correlations from random chance in literature data.

Main Results:

• The study presents a statistical approach capable of selecting true correlations, even when they are low-frequency associations.
• The e-LiSe application provides a tool for researchers to identify non-obvious biological associations.
• The method enhances the accuracy of literature mining for biomedical research.

Conclusions:

• The developed bio-linguistic statistical method effectively identifies significant biological correlations.
• The e-LiSe application offers a valuable tool for discovering understudied biological information.
• This approach improves the reliability of findings in bio-medical literature mining.