Journal: bioRxiv
Article Title: Accurate assignment of disease liability to genetic variants using only population data
doi: 10.1101/2021.04.19.440463
Figure Lengend Snippet: Data is jittered for display purposes only in Figures 1A-D ; precise allele counts and BayPR probabilities are provided in the Data Supplement. A Plot of the prevalence ratios (PR) of 22 of the most common pathogenic CFTR variants (black dots) according frequencies in affected (CFTR2) and healthy (gnomAD) individuals demonstrates expected deviation from neutrality (grey line). The designation of each of these variants as pathogenic by the American College of Medical Genetics (ACMG) remains unchanged since their initial interpretation in 2004. B Plot of the prevalence ratios of 135 pathogenic or likely pathogenic (P/LP; red dots) and 17 benign or likely benign (B/LB; green dots) variants interpreted by CFTR2 that have been observed at least once in gnomAD reveals two relatively distinct populations. C Plot of CFTR variants according allele count and BayPR probability of being disease-causing. All 313 variants have been interpreted by CFTR2 as CF-causing (pathogenic/likely pathogenic [P/LP]; n=296) or not CF-causing (benign/likely benign [B/LB]; n=17) using clinical and functional criteria. D BayPR probabilities of 313 CFTR variants stratified according to their functional consequence. E Receiver operating characteristic (ROC) curves were plotted for BayPR and ten in silico prediction algorithms using data from 101 CFTR missense variants. The reference standard are annotations in the CFTR2 database based on clinical and in vitro data. F Heatmap depicts estimated probability of a CFTR variant being associated with disease based on absolute counts of variants in the CFTR2 and gnomAD databases. The area in black represents a greater than 90% probability of being associated with CF and the area in light grey less than 10% probability. Importantly, this heatmap is only applicable to the CF data presented in this manuscript and does not apply to other genetic disorders. This heatmap was generated using the contour command in STATA IC 15, which uses thin plate splines for interpolation.
Article Snippet: Heatmaps were generated using the contour command in STATA IC 15, which uses thin plate splines for interpolation.
Techniques: Functional Assay, In Silico, In Vitro, Variant Assay, Generated
MathWorks Inc is a verified supplier 
![Data is jittered for display purposes only in Figures 1A-D ; precise allele counts and BayPR probabilities are provided in the Data Supplement. A Plot of the prevalence ratios (PR) of 22 of the most common pathogenic CFTR variants (black dots) according frequencies in affected (CFTR2) and healthy (gnomAD) individuals demonstrates expected deviation from neutrality (grey line). The designation of each of these variants as pathogenic by the American College of Medical Genetics (ACMG) remains unchanged since their initial interpretation in 2004. B Plot of the prevalence ratios of 135 pathogenic or likely pathogenic (P/LP; red dots) and 17 benign or likely benign (B/LB; green dots) variants interpreted by CFTR2 that have been observed at least once in gnomAD reveals two relatively distinct populations. C Plot of CFTR variants according allele count and BayPR probability of being disease-causing. All 313 variants have been interpreted by CFTR2 as CF-causing (pathogenic/likely pathogenic [P/LP]; n=296) or not CF-causing (benign/likely benign [B/LB]; n=17) using clinical and functional criteria. D BayPR probabilities of 313 CFTR variants stratified according to their functional consequence. E Receiver operating characteristic (ROC) curves were plotted for BayPR and ten in silico prediction algorithms using data from 101 CFTR missense variants. The reference standard are annotations in the CFTR2 database based on clinical and in vitro data. F Heatmap depicts estimated probability of a CFTR variant being associated with disease based on absolute counts of variants in the CFTR2 and gnomAD databases. The area in black represents a greater than 90% probability of being associated with CF and the area in light grey less than 10% probability. Importantly, this heatmap is only applicable to the CF data presented in this manuscript and does not apply to other genetic disorders. This heatmap was generated using the contour command in STATA IC 15, which uses thin plate splines for interpolation.](https://bio-rxiv-images-cdn.bioz.com/dois_ending_with_63/10__1101_slash_2021__04__19__440463/10__1101_slash_2021__04__19__440463___F1.large.jpg)