发帖
每发1个新帖
可以获得0.5个丁当奖励
回帖

【交流】在计算genetic power时的genotype relative risk

只看楼主
页码直达：
直达末页

楼主

oqpo

入门站友

1楼

这个帖子发布于9年零202天前，其中的信息可能已发生改变或有所发展。

问题已关闭悬赏丁当:10

在进行CASE-CONTROL研究时计算GENETIC POWER时，会要求提供genotype relative risk

使用genetic power calculator时，还有genotypic relative risk for Aa and AA。

但是在查阅文献时，让我困惑的是这个genotype relative risk似乎是人为设立的，而不是计算所得。

比如在这篇文章

Hoenicka J, Garrido E, Ponce G, Rodriguez-Jimenez R, Martinez I, Rubio G, Jimenez-Arriero MA, Palomo T. 2010.
Sexually dimorphic interaction between the DRD1 and COMT genes in schizophrenia.
Am J Med Genet Part B 153B: 948–954.

The power to detect association with the disorder in female and male dataset was estimated at 63%-69% and 73%-82%, respectively, with genotype relative risk of 1.8 at a nominal P=0.05 for minor alllele frequencies ranging from 0.15 to 0.38. For a genotype relative risk of 2, the power ranged from 76% to 82% in females and in excess of 85% to 92% in males.

不懂1.8和2是肿么来的。

邀请讨论

搜索

2011-07-12 11:15 浏览 : 5225 回复 : 4

投票
收藏 8
打赏
引用
分享
- 微信扫一扫
- 新浪微博
- 丁香客
- 复制网址
举报
- 广告宣传推广
- 政治敏感、违法虚假信息
- 恶意灌水、重复发帖
- 违规侵权、站友争执
- 附件异常、链接失效
- 其他

oqpo 编辑于 2011-07-12 15:58

• 误诊病例，等你来参与

persistentdog

常驻站友

2楼

不是吧，怎么会这么巧，我这两天也再研究这个方法。看的文献也是这篇做DRD1的文章。

relatvie risk，也很令人困惑。一般说病例对照研究无法获得RR值，所以一般用OR值代替，

我把文献男性群体的rs11746641和rs4867798 （marker B）两个位点的数据代入软件计算结果如下：

这个结果后一部分完全看不懂，请高手给指点下吧。

2011-07-13 14:18

投票
收藏 1
打赏
引用
分享
- 微信扫一扫
- 新浪微博
- 丁香客
- 复制网址
举报
- 广告宣传推广
- 政治敏感、违法虚假信息
- 恶意灌水、重复发帖
- 违规侵权、站友争执
- 附件异常、链接失效
- 其他

• 【分享】你所好奇的江苏三甲待遇！

z_lu_er

铁杆站友

3楼

As a measure of association between the marker and disease, Falk and Rubinstein (1987) defined the term haplotype relative risk (HRR) as an
odds ratio-the odds of the presence of a particular marker among either
of the two alleles transmitted to the affected child divided by the odds
that the marker is among either of the two nontransmitted alleles.

Ott (1989) further defined the statistical properties of the HRR for diseases with a recessive mode of inheritance.

One of his main findings was that, when random chance is ignored, the
HRR differs from unity, which implies an association, only when
disequilibrium exists between the disease and marker loci and the two
loci are linked. However, the HRR groups individuals who are homozygous
for the marker with heterozygous subjects and computes the odds ratio
contrasted to thebaseline group, which does not have the marker.Hence,
the HRR is implicitly a weighted average of two odds ratios: (1) an odds
ratio for those homozygous for the marker versus those homozygous
without the marker and (2) an odds ratio for those heterozygous for the
marker versus those homozygous without the marker. Our basic premise is
that, when assessing a candidate gene for association with disease, it
may at times

be more informative to examine each of these two odds ratios, or their corresponding relative risks, which we

will refer to as genotype relative risks (GRRs).

If cases are sampled from all those present in the general
population, the relative risks (odds ratios in a case–control study) for
specific alleles influencing complex diseases are expected to be modest
to small. For polymorphisms with allele frequencies >0.2,
the odds ratios are expected be in the range of 1.1–1.5; for allele
frequencies between 0.05 and 0.2, up to ~3.0. This is true
by definition, since a common variant with much larger relative risks
would result in a large attributable risk for that variant with respect
to the disease; in other words, the variant would explain a very large
proportion of the causality of the disease, which would make the
condition’s characteristics resemble a Mendelian rather than a complex
disorder. As a guideline, sample sizes of at least 1,000 cases and 1,000
controls are required to detect odds ratios ~1.5 in size with at least
80%power, but the required size of each individual study will depend on
whether (i) the analysis will also include case subgroups; (ii) the
analysis focuses on CGs with a limited number of independent tests or
GWAs with many thousands of tests; and (iii) there is an a priori
hypothesis to be tested relating to a polymorphism of known allele
frequency (e.g., in a replication study—see below). The expected effect
size to be detected in a study (and thus power) can sometimes be
increased using family history enrichment schemes for case sampling.

从上面的话来看, 在预测样本量大小或者在计算特定样本量的power的时候，所用到的GRR只是一个预设的值，也就是我们并不知道致病allele在病人和control中的频
率，所以无法明确的指导GRR。但是我们的前提假设是我们所研究的疾病是common disease (complex disease,
目前关联研究的主要对象)，也就是导致疾病发生的致病位点有很多，而且没有明显的对疾病发生贡献很大的位点，因此致病位点绝对不会像单基因病那样拥有很大
的GRR。按照上方标注黑色的区域的话来看，我们能选择的最大的GRR也就是3。我们不能选择很大的GRR来进行power的预测，否则power就不可
靠。最大的GRR是3的话，在显性遗传模式下，也就是两个纯和的致病allele同时存在时，相对于两个allele均不为致病allele是的OR值为
3。

再看power calculator的说明书：

Genotype Relative Risk: The definition of genotype relative risk (GRR) is depends on the disease model. If f0, f1, f2
are the probabilities of being affected for individuals with 0, 1, or 2 copies of the risk allele, then GRR is defined as follows:

MultiplicativeGRR = f1 / f0 = f2 / f1
AdditiveGRR = f1 / f0
DominantGRR = f1 / f0 = f2 / f0
RecessiveGRR = f2 / f0 = f2 / f1

这样我们就知道该如何设定我们的GRR了。
以上只是我稍微看了一些文献提取的，如有不妥之处，欢迎讨论。

2011-07-14 09:36

投票
收藏 3
打赏
引用
分享
- 微信扫一扫
- 新浪微博
- 丁香客
- 复制网址
举报
- 广告宣传推广
- 政治敏感、违法虚假信息
- 恶意灌水、重复发帖
- 违规侵权、站友争执
- 附件异常、链接失效
- 其他

• 首张“新冠患者分子全景图”

野骆驼

铁杆站友

4楼

persistentdog
不是吧，怎么会这么巧，我这两天也再研究这个方法。看的文献也是这篇做DRD1的文章。

relatvie risk，也很令人困惑。一般说病例对照研究无法获得RR值，所以一般用OR值代替，

我把文献男性群体的rs11746641和rs4867798 （marker B）两个位点的数据代入软件计算结果如下：

这个结果后一部分完全看不懂，请高手给指点下吧。

有看懂这结果的高人帮忙解读一下

2012-02-14 18:02

投票
收藏
打赏
引用
分享
- 微信扫一扫
- 新浪微博
- 丁香客
- 复制网址
举报
- 广告宣传推广
- 政治敏感、违法虚假信息
- 恶意灌水、重复发帖
- 违规侵权、站友争执
- 附件异常、链接失效
- 其他

• 回应：饶毅发文称“正式举报裴钢为通讯作者的文章涉嫌学术不端”

遗传发育

【交流】在计算genetic power时的genotype relative risk

提示

丁香园旗下网站

关于丁香园

官方链接

遗传发育

分享到：

微信扫一扫

【交流】在计算genetic power时的genotype relative risk

微信扫一扫

微信扫一扫

微信扫一扫

微信扫一扫

提示

丁香园旗下网站

关于丁香园

官方链接