Functional Prediction


This platform provides 13 published functional score including deleterious and conservation. These scores are used rankscore from dbNSFP database v2.9.2c. The rankscore is the ratio of the rank each score over the total number of each score in dbNSFP. The user can see the detail information in the boxplot figure, including the cutoff value, maximum, minimum, medimum, Q1 and Q3 value in each score.
 

 

Dr. Score

DNArails team provides one pathogenic score named DNArails Score (Dr. Score). We used the in-house machine learning algorithm to build the optimized classifier model that predicted all potential non-synonymous single-nucleotide variants (nsSNVs) in the human genome.
 



 

Out standing Performance of Dr. Score
Two benchmark independent test data were used to evaluate Dr. Score, which datasets are from published paper (PMCID: PMC4375422). Dr. Score achieved 0.93 AUC (the area under the ROC curve) in testing dataset 1 and 0.95 AUC in testing dataset 2.