東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

Evaluation of statistical methods to...

Nicodemus, Kristin Kay.

Linked to FindBook

Google Book

Amazon

博客來

Evaluation of statistical methods to detect epistasis and application to the DISC1 pathway and risk for schizophrenia.

Record Type:	Language materials, printed : Monograph/item
Title/Author:	Evaluation of statistical methods to detect epistasis and application to the DISC1 pathway and risk for schizophrenia./
Author:	Nicodemus, Kristin Kay.
Description:	139 p.
Notes:	Source: Dissertation Abstracts International, Volume: 68-04, Section: B, page: 2226.
Contained By:	Dissertation Abstracts International68-04B.
Subject:	Biology, Biostatistics. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3262479

Evaluation of statistical methods to detect epistasis and application to the DISC1 pathway and risk for schizophrenia.
Nicodemus, Kristin Kay.

Evaluation of statistical methods to detect epistasis and application to the DISC1 pathway and risk for schizophrenia. - 139 p.

Source: Dissertation Abstracts International, Volume: 68-04, Section: B, page: 2226.

Thesis (Ph.D.)--The Johns Hopkins University, 2007.

Disrupted-In-Schizophrenia (DISC1) was first described as a candidate gene for schizophrenia in an extended Scottish family, where a translocation within DISC1 segregated with psychiatric disorders. The etiology of schizophrenia is thought to be complex, including multiple susceptibility genes, environmental risk factors, and interactions between genes and genes and environment. To examine the ability of currently available methods to detect interaction, we conducted a simulation study assessing whether Random Forests, Monte Carlo Logic Regression (MCLR) and Generalized Boosted Regression (GBM) were able to detect several types of two-locus interaction models. A package for the R statistical computing environment (SH2IPS) was created for data simulation. We used the results from the simulation study to examine whether DISC1 and its partners epistatically influence risk for schizophrenia in a case-control study of 296 European American cases and 365 European American controls.Subjects--Topical Terms:

1018416
Biology, Biostatistics.

Evaluation of statistical methods to detect epistasis and application to the DISC1 pathway and risk for schizophrenia.
LDR:02903nam 2200253 a 45 001 945910
005 20110523
008 110523s2007 ||||||||||||||||| ||eng d
035 $a (UMI)AAI3262479
035 $a AAI3262479
040 $a UMI $c UMI
100 1 $a Nicodemus, Kristin Kay. $3 1269317
245 1 0 $a Evaluation of statistical methods to detect epistasis and application to the DISC1 pathway and risk for schizophrenia.
300 $a 139 p.
500 $a Source: Dissertation Abstracts International, Volume: 68-04, Section: B, page: 2226.
502 $a Thesis (Ph.D.)--The Johns Hopkins University, 2007.
520 $a Disrupted-In-Schizophrenia (DISC1) was first described as a candidate gene for schizophrenia in an extended Scottish family, where a translocation within DISC1 segregated with psychiatric disorders. The etiology of schizophrenia is thought to be complex, including multiple susceptibility genes, environmental risk factors, and interactions between genes and genes and environment. To examine the ability of currently available methods to detect interaction, we conducted a simulation study assessing whether Random Forests, Monte Carlo Logic Regression (MCLR) and Generalized Boosted Regression (GBM) were able to detect several types of two-locus interaction models. A package for the R statistical computing environment (SH2IPS) was created for data simulation. We used the results from the simulation study to examine whether DISC1 and its partners epistatically influence risk for schizophrenia in a case-control study of 296 European American cases and 365 European American controls.
520 $a Results from the simulation study suggested MCLR and GBM were the most promising methods for the detection of epistasis, even among highly correlated SNPs, and that RF is best to use for detection of SNPs involved in interaction in regions of moderate to low LD. The type of epistasis influenced the ability of all methods to detect epistasis; under a jointly dominant model all methods were more often able to detect causal variants than under other models such as jointly recessive. Evidence for epistasis between DISC1 and its protein interaction was observed, with one SNP in 3 genes (DISC1, LIS1, NDEL1) known to form a trimolecular complex in the brain scoring among the top 20 most influential SNPs. In addition, many SNPs in CITRON, also a DISC1 interaction partner, were detected as highly influential. Although statistical interaction can only suggest evidence of putative biologic interaction, results from the present study will be used to further test hypotheses at the biologic level, including using tools such as neuroimaging and postmortem mRNA expression profiling.
590 $a School code: 0098.
650 4 $a Biology, Biostatistics. $3 1018416
650 4 $a Health Sciences, Epidemiology. $3 1019544
690 $a 0308
690 $a 0766
710 2 $a The Johns Hopkins University. $3 1017431
773 0 $t Dissertation Abstracts International $g 68-04B.
790 $a 0098
791 $a Ph.D.
792 $a 2007
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3262479