Skip to main content

Table 1 Host prediction accuracies (%) for virus and host genomes from the data sets by Edwards et al. [14] and Galiez et al. [17]

From: Taxonomy-aware, sequence similarity ranking reliably predicts phage–host relationships

Data set

Method

Species

Genus

Family

Order

Class

Phylum

Edwards et al. [14]

WIsH

28

44

50

53

62

70

BLAST

43

59

71

78

87

96

Phirboa

43

59

71

78

87

96

Phirbo (+viruses)b

48

63

75

82

90

97

Galiez et al. [17]

WIsH

21

44

48

53

68

77

BLAST

31

53

62

68

88

95

Phirboa

31

53

62

68

88

95

Phirbo (+viruses)b

35

56

65

72

90

96

  1. The highest accuracies among the methods for each taxonomic level are in bold
  2. aPhirbo scores were calculated using rank-biased overlap (RBO) between BLAST lists containing prokaryotic sequences. Specifically, the BLAST database contained 2699 sequences of bacterial genomes in the Edwards et al. data set and 3780 sequences of bacterial and archaeal genomes in the Galiez et al. data set
  3. bPhirbo scores were calculated using RBO between BLAST lists containing both prokaryotic and viral sequences