Skip to main content

Table 5 Number of EE candidate genes prioritized in each version of each dataset. The prioritization method is discussed in detail in Additional file 1

From: Systematic noise degrades gene co-expression signals but can be corrected

Study

Raw

BC

BC+QN

RUV-random

Hawrylycz et al.

84

80

75

83

Miller et al.

80

82

84

86

Kang et al.

48

-a

33

62

Colantuoni et al.

33

42

43

44

Hernandez et al.

137

58

65

67b

  1. aPrioritization for the background-corrected Kang et al. dataset was not possible as no cutoff value for significant correlations could be defined
  2. bAs discussed earlier, we use the version of BC+RUV-random for this dataset