Limits...
Using DNA pools for genotyping trios.

Beckman KB, Abel KJ, Braun A, Halperin E - Nucleic Acids Res. (2006)

Bottom Line: The genotyping of mother-father-child trios is a very useful tool in disease association studies, as trios eliminate population stratification effects and increase the accuracy of haplotype inference.Unfortunately, the use of trios for association studies may reduce power, since it requires the genotyping of three individuals where only four independent haplotypes are involved.We demonstrate that the error rates in the genotype calls of the proposed protocol are comparable to those of standard genotyping techniques, although the cost is reduced considerably.

View Article: PubMed Central - PubMed

Affiliation: Children's Hospital Oakland Research Institute, 5700 Martin Luther King Jr Way, Oakland, CA 94609, USA.

ABSTRACT
The genotyping of mother-father-child trios is a very useful tool in disease association studies, as trios eliminate population stratification effects and increase the accuracy of haplotype inference. Unfortunately, the use of trios for association studies may reduce power, since it requires the genotyping of three individuals where only four independent haplotypes are involved. We describe here a method for genotyping a trio using two DNA pools, thus reducing the cost of genotyping trios to that of genotyping two individuals. Furthermore, we present extensions to the method that exploit the linkage disequilibrium structure to compensate for missing data and genotyping errors. We evaluated our method on trios from CEPH pedigree 66 of the Coriell Institute. We demonstrate that the error rates in the genotype calls of the proposed protocol are comparable to those of standard genotyping techniques, although the cost is reduced considerably. The approach described is generic and it can be applied to any genotyping platform that achieves a reasonable precision of allele frequency estimates from pools of two individuals. Using this approach, future trio-based association studies may be able to increase the sample size by 50% for the same cost and thereby increase the power to detect associations.

Show MeSH

Related in: MedlinePlus

The error rate of the parent–child pool genotyping versus standard genotyping as a function of increasing imprecision in allele frequency determination. All datasets are based on 200 SNPs in chromosome 22, taken from the HapMap CEU population, which includes 30 trios. For each allele frequency read (per SNP and pool), we added an error component, normally distributed with mean 0 and SD ranging from 0 to 10% (on the x-axis).
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC1636483&req=5

fig4: The error rate of the parent–child pool genotyping versus standard genotyping as a function of increasing imprecision in allele frequency determination. All datasets are based on 200 SNPs in chromosome 22, taken from the HapMap CEU population, which includes 30 trios. For each allele frequency read (per SNP and pool), we added an error component, normally distributed with mean 0 and SD ranging from 0 to 10% (on the x-axis).

Mentions: Current association studies only use pooled DNA data when the individual information is not needed, for instance, for the screening of potential linked SNPs as a first step of a multistage association study (8–10,18–20). In contrast to previous methods, the techniques presented here provide a new application for DNA pooling, namely the individual genotyping of trios. We have demonstrated that our method considerably increases the power of an association study when compared to a case–control based association study with an equivalent budget. The methodology used for this protocol is limited to trios and it assumes that the SNPs are biallelic and that there is no copy number variation. It may be possible to extend our methods to other scenarios in which more complex pedigree configurations are genotyped or even when unrelated individuals are genotyped. We emphasize that our choice of the genotyping platform for the evaluation of our method is arbitrary and is only used as a proof-of-principle. We expect the method to have similar performance on genotyping platforms that allows for reasonably accurate quantitative DNA pools measurements. In order to demonstrate this point, we used the HapMap CEU population to simulate scenarios in which the allele frequencies are read with errors. The simulations were performed over 200 SNPs in chromosome 22. The error was introduced by adding to the correct allele frequency a Gaussian distribution with mean 0 and standard deviation (SD) of allelic frequency estimation ranging from 0 to 10%. As can be seen in Figure 4, as long as the SD < 6%, the results are comparable to genotyping error rates. This is encouraging, since it implies that it may be possible to extend the approach to other genotyping platforms. In this study, empirical pool frequency determination using MALDI-TOF-based allelotyping across 12 SNPs resulted in a range of SDs (Min SD = 2.0%, Max SD = 10.0%, mean SD = 4.8 ± 2.3%, Supplementary Figure 3). For the current platform, at least, genotyping using parent–child pools in combination with Triophase should produce similar error rates to standard genotyping.


Using DNA pools for genotyping trios.

Beckman KB, Abel KJ, Braun A, Halperin E - Nucleic Acids Res. (2006)

The error rate of the parent–child pool genotyping versus standard genotyping as a function of increasing imprecision in allele frequency determination. All datasets are based on 200 SNPs in chromosome 22, taken from the HapMap CEU population, which includes 30 trios. For each allele frequency read (per SNP and pool), we added an error component, normally distributed with mean 0 and SD ranging from 0 to 10% (on the x-axis).
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC1636483&req=5

fig4: The error rate of the parent–child pool genotyping versus standard genotyping as a function of increasing imprecision in allele frequency determination. All datasets are based on 200 SNPs in chromosome 22, taken from the HapMap CEU population, which includes 30 trios. For each allele frequency read (per SNP and pool), we added an error component, normally distributed with mean 0 and SD ranging from 0 to 10% (on the x-axis).
Mentions: Current association studies only use pooled DNA data when the individual information is not needed, for instance, for the screening of potential linked SNPs as a first step of a multistage association study (8–10,18–20). In contrast to previous methods, the techniques presented here provide a new application for DNA pooling, namely the individual genotyping of trios. We have demonstrated that our method considerably increases the power of an association study when compared to a case–control based association study with an equivalent budget. The methodology used for this protocol is limited to trios and it assumes that the SNPs are biallelic and that there is no copy number variation. It may be possible to extend our methods to other scenarios in which more complex pedigree configurations are genotyped or even when unrelated individuals are genotyped. We emphasize that our choice of the genotyping platform for the evaluation of our method is arbitrary and is only used as a proof-of-principle. We expect the method to have similar performance on genotyping platforms that allows for reasonably accurate quantitative DNA pools measurements. In order to demonstrate this point, we used the HapMap CEU population to simulate scenarios in which the allele frequencies are read with errors. The simulations were performed over 200 SNPs in chromosome 22. The error was introduced by adding to the correct allele frequency a Gaussian distribution with mean 0 and standard deviation (SD) of allelic frequency estimation ranging from 0 to 10%. As can be seen in Figure 4, as long as the SD < 6%, the results are comparable to genotyping error rates. This is encouraging, since it implies that it may be possible to extend the approach to other genotyping platforms. In this study, empirical pool frequency determination using MALDI-TOF-based allelotyping across 12 SNPs resulted in a range of SDs (Min SD = 2.0%, Max SD = 10.0%, mean SD = 4.8 ± 2.3%, Supplementary Figure 3). For the current platform, at least, genotyping using parent–child pools in combination with Triophase should produce similar error rates to standard genotyping.

Bottom Line: The genotyping of mother-father-child trios is a very useful tool in disease association studies, as trios eliminate population stratification effects and increase the accuracy of haplotype inference.Unfortunately, the use of trios for association studies may reduce power, since it requires the genotyping of three individuals where only four independent haplotypes are involved.We demonstrate that the error rates in the genotype calls of the proposed protocol are comparable to those of standard genotyping techniques, although the cost is reduced considerably.

View Article: PubMed Central - PubMed

Affiliation: Children's Hospital Oakland Research Institute, 5700 Martin Luther King Jr Way, Oakland, CA 94609, USA.

ABSTRACT
The genotyping of mother-father-child trios is a very useful tool in disease association studies, as trios eliminate population stratification effects and increase the accuracy of haplotype inference. Unfortunately, the use of trios for association studies may reduce power, since it requires the genotyping of three individuals where only four independent haplotypes are involved. We describe here a method for genotyping a trio using two DNA pools, thus reducing the cost of genotyping trios to that of genotyping two individuals. Furthermore, we present extensions to the method that exploit the linkage disequilibrium structure to compensate for missing data and genotyping errors. We evaluated our method on trios from CEPH pedigree 66 of the Coriell Institute. We demonstrate that the error rates in the genotype calls of the proposed protocol are comparable to those of standard genotyping techniques, although the cost is reduced considerably. The approach described is generic and it can be applied to any genotyping platform that achieves a reasonable precision of allele frequency estimates from pools of two individuals. Using this approach, future trio-based association studies may be able to increase the sample size by 50% for the same cost and thereby increase the power to detect associations.

Show MeSH
Related in: MedlinePlus