|  | all predicted | positively correlated | negatively correlated |
---|
 | Hosts | targets | score | targets | score | targets | score |
---|
SCD I | 9 | 282 | 1.78 | 141 | 1.76 | 133 | 1.80 |
SCD II | 8 | 778 | 2.19 | 382 | 2.27 | 388 | 2.11 |
SG I | 13 | 1531 | 1.84 | 793 | 1.82 | 727 | 1.86 |
SG II | 21 | 1971 | 1.85 | 1013 | 1.90 | 942 | 1.80 |
SG III | 7 | 621 | 1.99 | 333 | 2.02 | 283 | 1.95 |
NO I | 10 | 873 | 1.81 | 351 | 1.81 | 512 | 1.80 |
NO II | 17 | 1286 | 2.19 | 706 | 2.22 | 567 | 2.16 |
- For each cluster we calculated the similarity score between host and all predicted target genes. Additionally, target genes were split up into positively and negatively correlated targets and similarity scores related to the host genes were calculated. The p-values determined by a comparison of functional GO similarities between host and predicted targets to randomly chosen sets of target genes of identical size were all significant after Bonferroni correction (all clusters p < 10-5). Independent from the sign of correlation, we determined a significantly higher similarity than expected for all analyzed clusters.