Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form

Examining polycystic ovary syndrome and its relevant biomarkers in Samoan women via machine learning-based phenotyping and genome-wide association studies

ERDOGAN-YILDIRIM, Zeynep (2022) Examining polycystic ovary syndrome and its relevant biomarkers in Samoan women via machine learning-based phenotyping and genome-wide association studies. Doctoral Dissertation, University of Pittsburgh. (Unpublished)

[img] PDF
Restricted to University of Pittsburgh users only until 30 August 2024.

Download (10MB) | Request a Copy


Polycystic ovary syndrome (PCOS) is a common endocrine disorder in reproductive-age women. It is highly heritable disorder with multifactorial and polygenic etiology. Genome-wide studies in various populations revealed several significant loci. Yet, the underlying mechanisms remain to be explored. Because studies in genetically isolated populations are valuable resource to gain novel biological insights, I investigated the genetic factors of PCOS and relevant biomarkers in Samoan women from the Soifua Manuia Study.
The first aim required phenotyping of study participants since no clinical diagnosis of PCOS was available. Taking advantage of supervised machine learning methods, several models were trained, tested and validated on European cohorts with known outcomes. The best performing model was ranger, a random forest algorithm, which I used to classify individuals into high-risk PCOS cases and low-risk PCOS controls.
In the second aim, I explored genetic factors associated with PCOS on 584 women from two independent Samoan samples via genome-wide association meta-analysis. Five loci showed suggestive association: TBX3, ERCC6L2, KRT4, C9orf163/SEC16A and CA8. Although none of the loci have known direct causal role in PCOS, they were either involved in gonadogenesis, were differentially expressed in adipose tissue of women with PCOS or were closely located to known PCOS loci.
The third aim interrogated genetic factors regulating the levels of cardinal biomarkers of PCOS – total testosterone (TT), free androgen index (FAI), sex hormone-binding globulin (SHBG) and anti-Müllerian hormone (AMH) – in a meta-analysis of 1103 Samoan women. I observed that rs1220069967 in ABO was genome-wide significant for both TT and FAI and is in high linkage disequilibrium with rs494242 in ABO that demonstrated association with TT levels in Europeans. Additional suggestive association signals were found for TT (N = 11), FAI (N = 12), SHBG (N = 7) and AMH (N = 14).
These promising preliminary findings provide new insights into the genetic underpinnings of PCOS and its biomarkers in Samoans. Their validation through additional genetic and functional follow-up studies is required. These findings expand our knowledge and may promote public health through the discovery of biomarkers for efficient screening and early diagnosis along with novel therapeutic targets for PCOS.


Social Networking:
Share |


Item Type: University of Pittsburgh ETD
Status: Unpublished
CreatorsEmailPitt UsernameORCID
ERDOGAN-YILDIRIM, Zeynepzee5@pitt.eduzee50000-0001-5349-6562
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Committee ChairMinster, Ryan Lrminster@pitt.edurminster
Committee MemberMassart, Mylynda Bmassartmb@upmc.edumbm66
Committee MemberTalbott, Evelyneot1@pitt.edueot1
Committee MemberWeeks, Daniel Eweeks@pitt.eduweeks
Date: 30 August 2022
Date Type: Publication
Defense Date: 21 July 2022
Approval Date: 30 August 2022
Submission Date: 30 August 2022
Access Restriction: 2 year -- Restrict access to University of Pittsburgh for a period of 2 years.
Number of Pages: 204
Institution: University of Pittsburgh
Schools and Programs: School of Public Health > Human Genetics
Degree: PhD - Doctor of Philosophy
Thesis Type: Doctoral Dissertation
Refereed: Yes
Uncontrolled Keywords: PCOS, Samoa, GWAS, genetic association study
Date Deposited: 30 Aug 2022 14:26
Last Modified: 30 Aug 2022 14:26


Monthly Views for the past 3 years

Plum Analytics

Actions (login required)

View Item View Item