Skip to content

Latest commit

 

History

History
11 lines (7 loc) · 794 Bytes

Performance Evaluation.md

File metadata and controls

11 lines (7 loc) · 794 Bytes

Performance Evaluation

  • Prediction precision

    • If a predicted poly(A) site is within a given distance (e.g., 50 bp, 100bp, 150bp) of any annotated poly(A) site (all_PA_hg19.txt), then it is considered as a true positive (TP), otherwise is a false positive (FP).

  • Sensitivity

    • A cutoff of 50 bp was used to determine whether a predicted poly(A) site is a true positive or not. Top 5000 to 30,000 annotated poly(A) sites according to the supported number of reads were chosen as the reference poly(A) sites.

  • Demo data

    • Prediction results for the human MAQC brain data from each benchmarked tool were stored in the result_data folder. These files are then processed and used for performance evaluation.