Figure 4.

Performance evaluation of SELINA and existing annotation tools

(A and B) The prediction accuracy and MacroF1 using datasets from the same sequencing platform (A) and multiple sequencing platforms (B) as reference. The number in each cell represents the average performance of all tested datasets within one tissue. The top row shows the average performance of all tissues for each method. The number of datasets included in each tissue is listed behind the tissue names.

(C) Mean training time (3 repeats) for increasing training cell numbers.

(D) Mean testing time (3 repeats) for increasing testing cell numbers.