Skip to main content

Table 4 Evaluation of the accuracy of error-correction methods

From: Benchmarking of computational error-correction methods for next-generation sequencing data

  Metric name Metric formula
a. Precision \( =\Big\{{\displaystyle \begin{array}{rr}\frac{\mathrm{TP}}{\mathrm{TP}+\mathrm{FP}+{\mathrm{FP}}_{\mathrm{INDEL}}},& \mathrm{TP}+\mathrm{FP}+{\mathrm{FP}}_{\mathrm{INDEL}}>0\\ {}0,& \mathrm{TP}+\mathrm{FP}+{\mathrm{FP}}_{\mathrm{INDEL}}=0\end{array}} \)
b. Sensitivity \( =\Big\{{\displaystyle \begin{array}{rr}\frac{\mathrm{TP}}{\mathrm{TP}+\mathrm{FN}},& \mathrm{TP}+\mathrm{FN}>0\\ {}0,& \mathrm{TP}+\mathrm{FN}=0\end{array}} \)
c. Gain \( =\Big\{{\displaystyle \begin{array}{rr}\frac{\mathrm{TP}-\left(\mathrm{FP}+{\mathrm{FP}}_{\mathrm{INDEL}}\right)}{\mathrm{TP}+\mathrm{FN}},& \mathrm{TP}+\mathrm{FN}>0\\ {}0,& \mathrm{TP}+\mathrm{FN}=0\end{array}} \)
d. Trim percent \( =\Big\{{\displaystyle \begin{array}{rr}\frac{{\mathrm{TP}}_{\mathrm{TRIM}}+{\mathrm{FP}}_{\mathrm{TRIM}}}{\mathrm{TotalBases}},& \mathrm{TotalBases}>0\\ {}0,& \mathrm{TotalBases}=0\end{array}} \)
e. Trim efficiency \( =\Big\{{\displaystyle \begin{array}{rr}\frac{{\mathrm{TP}}_{\mathrm{TRIM}}}{{\mathrm{TP}}_{\mathrm{TRIM}}+F{P}_{\mathrm{TRIM}}},& {\mathrm{TP}}_{\mathrm{TRIM}}+{\mathrm{FP}}_{\mathrm{TRIM}}>0\\ {}0,& {\mathrm{TP}}_{\mathrm{TRIM}}+{\mathrm{FB}}_{\mathrm{TRIM}}=0\end{array}} \)
  1. a. Precision evaluates the proportion of proper corrections among the total number of performed corrections. INDEL refers to insertion/deletion polymorphism. b. Sensitivity evaluates the proportion of fixed errors among all existing errors in the data. c. Gain represents whether an algorithm is producing an overall benefit (more TP then FP) or is having a negative effect (more FP then TP). Values ranging from 1.0 to, but not including, 0.0 represent a benefit; 0.0 is neutral; and less than 0.0 is considered a negative effect. d. Trim percent is the proportion of nucleotides trimmed out of all nucleotides analyzed. e. Trim efficiency is the proportion of trimmed bases from the tool that were considered to be TP trimming