Limits...
Influence analysis in quantitative trait loci detection.

Dou X, Kuriki S, Maeno A, Takada T, Shiroishi T - Biom J (2014)

Bottom Line: We derive general formulas of influence functions for profile likelihoods and introduce them into two standard quantitative trait locus detection methods-the interval mapping method and single marker analysis.These methods are shown useful in the influence analysis of a real dataset of an experimental population from an F2 mouse cross.By receiver operating characteristic analysis, we confirm that the proposed methods show better performance than existing diagnostics.

View Article: PubMed Central - PubMed

Affiliation: The Institute of Statistical Mathematics, Research Organization of Information and Systems, 10-3 Midori-cho, Tachikawa, Tokyo, 190-8562, Japan.

Show MeSH

Related in: MedlinePlus

Simulation results for Section 6.1. Comparison of ROC curves. EIFC: linear combination of EIFs, QEIF: quadratic EIF, Cook: Cook's D, Resid: standardized residual r. Distribution of outliers, εi, is (A) N(0,(2σ)2), (B) N(0,(3σ)2), (C) t3 with scale 2σ, and (D) Cauchy with scale 2σ.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4282084&req=5

fig05: Simulation results for Section 6.1. Comparison of ROC curves. EIFC: linear combination of EIFs, QEIF: quadratic EIF, Cook: Cook's D, Resid: standardized residual r. Distribution of outliers, εi, is (A) N(0,(2σ)2), (B) N(0,(3σ)2), (C) t3 with scale 2σ, and (D) Cauchy with scale 2σ.

Mentions: The ROC curves of the four indicators are shown in Figure 5 based on 1000 replicates. For the varying thresholds, the average rates of correctly classifying the true influential cases (detection rates), and the average rates of misclassifying the normal cases as influential cases (false positive rates) are plotted. As expected, outliers with larger σ2 have larger absolute EIFs, and are thus more easily detected. In all panels, the EIFC has the largest area under its ROC curve and hence the best average performance. As the second-best method, QEIF is shown useful when the target shape cannot be fully specified.


Influence analysis in quantitative trait loci detection.

Dou X, Kuriki S, Maeno A, Takada T, Shiroishi T - Biom J (2014)

Simulation results for Section 6.1. Comparison of ROC curves. EIFC: linear combination of EIFs, QEIF: quadratic EIF, Cook: Cook's D, Resid: standardized residual r. Distribution of outliers, εi, is (A) N(0,(2σ)2), (B) N(0,(3σ)2), (C) t3 with scale 2σ, and (D) Cauchy with scale 2σ.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4282084&req=5

fig05: Simulation results for Section 6.1. Comparison of ROC curves. EIFC: linear combination of EIFs, QEIF: quadratic EIF, Cook: Cook's D, Resid: standardized residual r. Distribution of outliers, εi, is (A) N(0,(2σ)2), (B) N(0,(3σ)2), (C) t3 with scale 2σ, and (D) Cauchy with scale 2σ.
Mentions: The ROC curves of the four indicators are shown in Figure 5 based on 1000 replicates. For the varying thresholds, the average rates of correctly classifying the true influential cases (detection rates), and the average rates of misclassifying the normal cases as influential cases (false positive rates) are plotted. As expected, outliers with larger σ2 have larger absolute EIFs, and are thus more easily detected. In all panels, the EIFC has the largest area under its ROC curve and hence the best average performance. As the second-best method, QEIF is shown useful when the target shape cannot be fully specified.

Bottom Line: We derive general formulas of influence functions for profile likelihoods and introduce them into two standard quantitative trait locus detection methods-the interval mapping method and single marker analysis.These methods are shown useful in the influence analysis of a real dataset of an experimental population from an F2 mouse cross.By receiver operating characteristic analysis, we confirm that the proposed methods show better performance than existing diagnostics.

View Article: PubMed Central - PubMed

Affiliation: The Institute of Statistical Mathematics, Research Organization of Information and Systems, 10-3 Midori-cho, Tachikawa, Tokyo, 190-8562, Japan.

Show MeSH
Related in: MedlinePlus