Making Sense of Messy Data: Classifying Metadata Variables in Public Transcriptomic (RNA) Data

Authors

Grace S. Brown, Brigham Young University - ProvoFollow
Tolulope Akinbo, Brigham Young University - Provo
Stephen Piccolo, Brigham Young University - ProvoFollow

Files

Keywords

Machine Learning, Prediction, Ontology

Abstract

US researchers are required to deposit data in public repositories so other researchers can validate their findings and reuse their data for new studies. There are > 100,000 transcriptomic (RNA) datasets in the public domain. However, it is of little use without metadata (details about individual samples e.g. sex, race). Having quality metadata provides many benefits including

• controlling for confounding factors.
• improving precision medicine.

Research has suggested

• males are studied more than females.
• people of European descent are studied most.

We need to make representation more equitable so genetic research will be relevant to all people. The first step is characterizing current subgroup representation in more detail. But classifying datasets by their metadata is time intensive and requires manual curation because metadata are described inconsistently across studies.

BYU ScholarsArchive Citation

Brown, Grace S.; Akinbo, Tolulope; and Piccolo, Stephen, "Making Sense of Messy Data: Classifying Metadata Variables in Public Transcriptomic (RNA) Data" (2024). Library/Life Sciences Undergraduate Poster Competition 2024. 58.
https://scholarsarchive.byu.edu/library_studentposters_2024/58

Document Type

Poster

Publication Date

2024-03-21

Language

English

College

Life Sciences

Department

Biology

University Standing at Time of Publication

Senior

Copyright Use Information

https://lib.byu.edu/about/copyright/