TY - GEN
T1 - Enrich rare disease phenotypic characterizations via a graph convolutional network based recommendation system
AU - Shen, Feichen
AU - Wen, Andrew
AU - Liu, Hongfang
N1 - Funding Information:
This work has been supported by the National Institute of Health (NIH) grant U01TR0062-1.
Publisher Copyright:
© 2020 IEEE.
PY - 2020/7
Y1 - 2020/7
N2 - Nowadays, there exist more than 300 million patients affected by about 7,000 rare disease all over the world, which comprises 3.5% to 5.9% of the global population. 40% of rare disease patients are diagnosed incorrectly before reaching a final diagnosis, of which 25% spend between 5 to 30 years on a chaotic journey through numerous referrals, investigations, and disease evolutions from early symptoms to a confirmatory diagnosis of their disease. Phenotypes are defined as observable characteristics and clinical traits of diseases and organisms. A significant lack of knowledge and insufficient characterization of the longitudinal phenotypic information of many rare diseases is a significant contributor to the continued existence of such diagnostic odyssey. In this study, to largely detect longitudinal phenotypic characterizations for rare disease, we formulated the problem of enriching rare disease phenotypic sets as a phenotype recommendation task and applied the graph convolutional network along with biomedical knowledge graph over Mayo Clinic electronic health records to achieve the goal.
AB - Nowadays, there exist more than 300 million patients affected by about 7,000 rare disease all over the world, which comprises 3.5% to 5.9% of the global population. 40% of rare disease patients are diagnosed incorrectly before reaching a final diagnosis, of which 25% spend between 5 to 30 years on a chaotic journey through numerous referrals, investigations, and disease evolutions from early symptoms to a confirmatory diagnosis of their disease. Phenotypes are defined as observable characteristics and clinical traits of diseases and organisms. A significant lack of knowledge and insufficient characterization of the longitudinal phenotypic information of many rare diseases is a significant contributor to the continued existence of such diagnostic odyssey. In this study, to largely detect longitudinal phenotypic characterizations for rare disease, we formulated the problem of enriching rare disease phenotypic sets as a phenotype recommendation task and applied the graph convolutional network along with biomedical knowledge graph over Mayo Clinic electronic health records to achieve the goal.
KW - Biomedical knowledge graph
KW - Graph convolutional network
KW - Phenotypic characterization
KW - Rare disease
KW - Recommendation system
UR - http://www.scopus.com/inward/record.url?scp=85091133890&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85091133890&partnerID=8YFLogxK
U2 - 10.1109/CBMS49503.2020.00015
DO - 10.1109/CBMS49503.2020.00015
M3 - Conference contribution
AN - SCOPUS:85091133890
T3 - Proceedings - IEEE Symposium on Computer-Based Medical Systems
SP - 37
EP - 40
BT - Proceedings - 2020 IEEE 33rd International Symposium on Computer-Based Medical Systems, CBMS 2020
A2 - de Herrera, Alba Garcia Seco
A2 - Rodriguez Gonzalez, Alejandro
A2 - Santosh, KC
A2 - Temesgen, Zelalem
A2 - Kane, Bridget
A2 - Soda, Paolo
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 33rd IEEE International Symposium on Computer-Based Medical Systems, CBMS 2020
Y2 - 28 July 2020 through 30 July 2020
ER -