Additional file 3 of A benchmark study on error-correction by read-pairing and tag-clustering in amplicon-based deep sequencing

Figure S3. Error rate correlation among different error-correction schemes. (a) Linear regression between true mutations and different error-correction methods. The model y∟x+a was adapted to do regression. Every dot represents a position on the target sequence and the values on x-axis and y-axis represent error rates of combined consensus and certain consensus, respectively. Colored lines are regression result. (b) Barplot of the intercepts a from the linear regression. Error bar is standard error. The colors represents different error-correction schemes, which are labeled in the graph. (EPS 477 kb