Skip to main content

Table 2 tRigon runtime based on data frame size and computational setting

From: tRigon: an R package and Shiny App for integrative (path-)omics data analysis

Task

Setting A

Setting B

Processing data

 Small

 Medium

39,280 ms

20,190 ms

 Large

283,870 ms

114,790 ms

Loading data

 Small

1520 ms

1290 ms

 Medium

5370 ms

4460 ms

 Large

47,550 ms

29,740 ms

Summary statistics (1 feature)

 Small

650 ms

230 ms

 Medium

1320 ms

1200 ms

 Large

13,590 ms

4020 ms

Distribution plots (1 feature, violin plot, no log-scale)

 Small

2410 ms

1290 ms

 Medium

1980 ms

1220 ms

 Large

5780 ms

3280 ms

Statistical tests (1 feature, pairwise-Wilcoxon test)

 Small

170 ms

70 ms

 Medium

290 ms

80 ms

 Large

1180 ms

570 ms

k-means clustering (6 features, 4 cluster, no groups)

 Small

1450 ms

840 ms

 Medium

2970 ms

730 ms

 Large

5590 ms

1930 ms

Feature importance (classification, 6 features, random forest)

 Small

2000 ms

1040 ms

 Medium

6090 ms

2990 ms

 Large

43,980 ms

22,870 ms

Correlation matrix (multiple corr., 6 features, no subgroups)

 Small

1220 ms

510 ms

 Medium

1460 ms

460 ms

 Large

1570 ms

710 ms

  1. All tasks were monitored with three dataframes (small: 281 rows, 36 columns, 55.9 KB size; medium: 211,287 rows, 53 columns, 47.7 MB size; large: 2,385,605 rows, 42 columns, 228 MB size) and in two computational settings (A: Intel Pentium CPU 1.60 GHz, 8 GB RAM; B: Intel Xeon Gold 6128 CPU 3.40 GHz, 128 GB RAM). The small dataframe is a medical dataset (which cannot be processed) while the medium and large dataframes are pathomics datasets
  2. ms milliseconds, log logarithmic, corr. correlation