Skip to main content

Table 1 Tabulates the total number of ontologies that were found for a given number of clusters for each dataset, as well as the number of significantly enriched ontologies.

From: Assessing and selecting gene expression signals based upon the quality of the measured dynamics

   Edge Only LOOCV + EDGE
Gene Set Clusters Significant Ontologies All Ontologies Significant Ontologies All Ontologies
GDS253 2 42 1696 136 1049
  3 133 1696 202 1049
  4 183 1696 288 1049
  5 222 1696 304 1049
  6 261 1696 352 1049
  7 330 1696 392 1049
  8 441 1696 438 1049
  9 455 1696 491 1049
  10 470 1696 499 1049
  11 497 1696 517 1049
  12 537 1696 540 1049
  13 584 1696 544 1049
  14 598 1696 558 1049
  15 620 1696 585 1049
  16 642 1696 593 1049
  17 673 1696 612 1049
  18 721 1696 619 1049
  19 733 1696 632 1049
GDS599 2 78 1292 96 140
  3 129 1292 109 140
  4 172 1292 117 140
  5 219 1292 121 140
  6 282 1292 121 140
  7 344 1292 122 140
  8 420 1292 128 140
  9 444 1292 128 140
  10 456 1292 131 140
  11 484 1292 131 140
  12 544 1292 131 140
  13 569 1292 131 140
  14 600 1292 132 140
  15 632 1292 133 140
  16 644 1292 133 140
  17 658 1292 133 140
  18 684 1292 133 140
  19 708 1292 134 140
GDS972 2 189 2372 98 1696
  3 211 2372 133 1696
  4 230 2372 183 1696
  5 265 2372 222 1696
  6 310 2372 261 1696
  7 330 2372 330 1696
  8 374 2372 411 1696
  9 413 2372 455 1696
  10 450 2372 470 1696
  11 471 2372 497 1696
  12 521 2372 537 1696
  13 558 2372 584 1696
  14 582 2372 598 1696
  15 609 2372 620 1696
  16 656 2372 642 1696
  17 707 2372 673 1696
  18 714 2372 721 1696
  19 773 2372 733 1696
  1. In all of the cases, we see that after running the LOOCV filter as well as the selection via EDGE, the total number of ontologies that have been selected is lower, whereas the number of statistically significantly enriched ontologies remains relatively constant. Thus, the improvement in the total number of ontologies appears to be due primarily to the removal of genes that do not show significant co-functionality.
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us