Disclaimer: Supplementary materials have been peer-reviewed but not copyedited.


Fig. SI. Box plot of the findings on the 20 items of the original questionnaire in Delphi round 1. Relative probabilities (RP) are presented on a log scale.

Fig. SII. Correlation matrix for the 20 items of the original questionnaire* in the first Delphi round. The shape of an ellipse indicates the magnitude of the correlation: circle = no correlation; line = perfect correlation. The color of an ellipse indicates the direction of the correlation: blue = positive correlation; red = negative correlation.

Fig. SIII. For each of the seven traits, percentage of SNPs for which the specific type of prior knowledge is available: comparison between “true” and random SNPs. In the first graph, results are pooled and percentages are calculated as weighted average across the seven traits. RP, pooled relative probability (and 95% confidence interval) from fixed effect meta-analysis. Relative probabilities correspond to odds ratios.

Table SI. Change in experts’ opinions from Delphi round 1 (R1) to 2 (R2). Expert 9 did not participate in R 2. Example of how the questions (Items, “I_”) were formulated: “How many times more likely to be associated with the phenotype is an SNP in a functional protein domain compared with a SNP which is not?”

Table SII. Comparison in the formulation of questions to the experts vs. queries for the bioinformatics tool

Table SIII. Relative probabilities of association for each type of prior knowledge based on experts’ opinions. Relative probabilities (“how many time more likely”) are reported together with 95% confidence intervals. RP, relative probability; 95%CI, 95% confidence intervals. Sensitivity analysis performed without outlier, Expert 4

Table SIV. Relative probabilities of association for each type of prior knowledge based on empirical evidence. Relative probabilities are reported with 95% confidence intervals. GFR, glomerular filtration rate; BMI, body mass index; RP, relative probability; NA, not available

Table SV. Results of the sensitivity analyses performed for three of seven traits (renal dysfunction, Crohn's disease, coronary artery disease) by extending the list of UMLS CUI terms to cover “closely related” phenotypes as in the questions to experts, for all questions referring to some previous links with the phenotype. Results from experts’ opinion are reported for comparison

Table SVI. Dependence between questions in the empirical evidence: Pearson correlation (phi coefficients) estimated in the 7,000 random SNPs. Marked with * are statistically significant correlations, after adjustment for multiple testing (P-value threshold after Bonferroni correction: 0.05/98 = 5.1 × 10−4). Highlighted in gray are correlations with coefficients ≥ 0.25. ME, mutually exclusive

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.