Display Settings:

Items per page
We are sorry, but NCBI web applications do not support your browser and may not function properly. More information

Results: 7

1.
Figure 5

Figure 5. From: Gene annotation and network inference by phylogenetic profiling.

Phylogenetic Profiles of the 11-member cluster (Figure 5) of orthologs across 66 genomes uncovered by CE. Green represents absence and red, presence of an ortholog.

Jie Wu, et al. BMC Bioinformatics. 2006;7:80-80.
2.
Figure 2

Figure 2. From: Gene annotation and network inference by phylogenetic profiling.

PPV as a function of GO depth for CE and SGA decision rules, using a correlation threshold of C* = 0.40 (p* = 10-8). The predictive coverage is approximately 25%.

Jie Wu, et al. BMC Bioinformatics. 2006;7:80-80.
3.
Figure 1

Figure 1. From: Gene annotation and network inference by phylogenetic profiling.

Specificity and positive predictive value as a function of predictive coverage for SGA and CE decision rules. Coverage is a function of correlation threshold, C*.

Jie Wu, et al. BMC Bioinformatics. 2006;7:80-80.
4.
Figure 6

Figure 6. From: Gene annotation and network inference by phylogenetic profiling.

Number of clusters (size >= 3) as a function of C*. Shaded area, i.e. C*>0.7 where SGA and CE have relatively small PPV difference, is used to extract functional and evolutionary modules.

Jie Wu, et al. BMC Bioinformatics. 2006;7:80-80.
5.
Figure 4

Figure 4. From: Gene annotation and network inference by phylogenetic profiling.

Expanded view of true positive and predicted clusters (Figure 4) in functional category H, showing two strikingly dense clusters of size 11 and 7. The elements in the larger cluster all participate in Porphyrin and chlorophyll metabolism (KEGG: 00860), which is a subset of category H (coenzyme metabolism).

Jie Wu, et al. BMC Bioinformatics. 2006;7:80-80.
6.
Figure 7

Figure 7. From: Gene annotation and network inference by phylogenetic profiling.

Evolutionarily conserved densely connected clusters. Edge coding: black, C* = 0.91 (p* = 10-18); yellow, C* = 0.81 (p* = 10-16); blue, C* = 0.71 (p* = 10-14). Green nodes correctly annotated; red nodes unannotated. (a) One of 4 cliques uncovered at C* = 0.91. All genes are known and are in the flagella assembly pathway. (b) One of 9 clusters at C* = 0.81, ranging in size from 4 to 13 nodes. In the cluster shown, all genes are annotated to the KEGG histidine metabolism pathway and to the COG amino acid metabolism and transport category. (c) – (e) are examples of mixed annotated-unannotated clusters, with the annotated sets homogeneous in function. Lower bounds on PPV for assigned functions are 79% (C* = 0.71) and 89% (C* = 0.81).

Jie Wu, et al. BMC Bioinformatics. 2006;7:80-80.
7.
Figure 3

Figure 3. From: Gene annotation and network inference by phylogenetic profiling.

An all against all VisANT screen shot, at C* = 0.55, of the 4286 orthologs in the COG database. 926 genes (677 annotated; 249 unannotated) are linked to at least one annotated gene. Each gene is unambiguously assigned to a unique COG functional category. Of the 677 annotated genes, 463 are correctly assigned; In total 1843 out of 4286 orthologs are unannotated in the COG classification. (A) Complete 926 gene network. (B) meta-network of genes from (A). Each group represents a set of genes allocated to a COG functional category using CE. (C) Detail of functional category H, coenzyme metabolism. (D). Of the 926 linked genes, 82 are in category H. 62 of them are true positives (green) and 8 are predictions (red). The remaining 12 are annotated in a different functional category and are therefore putative false positives. The minimum PPV for category H is therefore 62/74 = 0.84 the averaged PPV for all categories is estimated to be 68% from all annotated genes. Refer to the COG web site for definitions of categories.

Jie Wu, et al. BMC Bioinformatics. 2006;7:80-80.

Display Settings:

Items per page

Supplemental Content

Recent activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...
Write to the Help Desk