Skip to main content

Table 1 Results of the KDVS for Prostate cancer study

From: Knowledge Driven Variable Selection (KDVS) – a new approach to enrichment analysis of gene signatures obtained from high–throughput data

Kind Enrichment KDVS
Lists PTvsM PTvsN PTvsM PTvsM PTvsN PTvsN
    MF BP MF BP
Discr. GO Terms 60 120 1115 2242 320 689
Discr. Genes 59 389 3619 4457 3118 3271
Comm. GO Terms 27 61 375 1000 158 378
Comm. Genes 7 51 418 504 334 371
Bench. Cov. GO Terms 1% 2% 46% 41% 19% 16%
Bench. Cov. Genes 1% 6% 49% 59% 39% 44%
  1. In the first two rows, the table shows the discriminant gene and GO term lists identified by enrichment–based approach and KDVS for each classification problem solved: Primary Tumor (PT) versus Metastasis (M) and Primary Tumor vs Normal (N). For KDVS, in addition, each classification task was addressed two times, based on the GO domain utilized, Molecular Function (MF) or Biological Process (BP). Discr. stands for discriminant.
  2. In the next two rows, the table shows the gene and GO term lists obtained as an intersection between respective discriminant lists and benchmark lists, built for prostate cancer. Comm. means common.
  3. In the last two rows, the table shows the coverage of benchmark gene and GO term lists with respective discriminant lists. For KDVS, the coverage was calculated only for benchmark MF and BP terms, respectively, according to GO domain utilized in classification task. Benchmark gene list consists of 851 elements. Benchmark GO term list consists of 2437 BP terms and 824 MF terms, 3593 terms in total. Bench. cov. means benchmark coverage.
Back to article page

Source Code for Biology and Medicine

ISSN: 1751-0473