Limits...
Assessment of subnetwork detection methods for breast cancer.

Jiang B, Gribskov M - Cancer Inform (2014)

Bottom Line: Here, we compare the results of eight methods: simulated annealing-based jActiveModules, greedy search-based jActiveModules, DEGAS, BioNet, NetBox, ClustEx, OptDis, and NetWalker.While the number of genes/proteins and protein interactions detected by the eight methods vary widely, a core set of 60 genes and 50 interactions was found to be shared by the subnetworks identified by five or more of the methods.Within the core set, 12 genes were found to be known breast cancer genes.

View Article: PubMed Central - PubMed

Affiliation: Department of Biological Sciences, Purdue University, West Lafayette, IN, USA.

ABSTRACT
Subnetwork detection is often used with differential expression analysis to identify modules or pathways associated with a disease or condition. Many computational methods are available for subnetwork analysis. Here, we compare the results of eight methods: simulated annealing-based jActiveModules, greedy search-based jActiveModules, DEGAS, BioNet, NetBox, ClustEx, OptDis, and NetWalker. These methods represent distinctly different computational strategies and are among the most widely used. Each of these methods was used to analyze gene expression data consisting of paired tumor and normal samples from 50 breast cancer patients. While the number of genes/proteins and protein interactions detected by the eight methods vary widely, a core set of 60 genes and 50 interactions was found to be shared by the subnetworks identified by five or more of the methods. Within the core set, 12 genes were found to be known breast cancer genes.

No MeSH data available.


Related in: MedlinePlus

Number of methods detecting genes and interactions in subnetworks. Histograms of the number of genes (A) and interaction counts (B) versus the number of methods that detect them. (A) All genes denote the 7,369 genes in the HPRD network. Breast cancer genes are the 462 genes found by KOBAS in multiple disease databases. Both the gene counts are scaled to [0, 1] by dividing by the maximum count. The percentage of breast cancer genes is the breast cancer gene count divided by the count of all the genes in each category (genes found by a certain number of methods). (B) All interactions denote the 28,571 interactions in the HPRD network. Breast cancer pathways are the 2,058 interactions found by KOBAS in multiple pathways databases. Both the interaction counts are scaled to [0, 1] by dividing by the maximum count. The percentage of breast cancer pathways is the interaction count in breast cancer pathways divided by the total interaction count in each category.
© Copyright Policy - open-access
Related In: Results  -  Collection


getmorefigures.php?uid=PMC4256043&req=5

f5-cin-suppl.6-2014-015: Number of methods detecting genes and interactions in subnetworks. Histograms of the number of genes (A) and interaction counts (B) versus the number of methods that detect them. (A) All genes denote the 7,369 genes in the HPRD network. Breast cancer genes are the 462 genes found by KOBAS in multiple disease databases. Both the gene counts are scaled to [0, 1] by dividing by the maximum count. The percentage of breast cancer genes is the breast cancer gene count divided by the count of all the genes in each category (genes found by a certain number of methods). (B) All interactions denote the 28,571 interactions in the HPRD network. Breast cancer pathways are the 2,058 interactions found by KOBAS in multiple pathways databases. Both the interaction counts are scaled to [0, 1] by dividing by the maximum count. The percentage of breast cancer pathways is the interaction count in breast cancer pathways divided by the total interaction count in each category.

Mentions: Then we used the list of true breast cancer genes to investigate if cancer-related genes are more likely to be detected by multiple methods. The distribution of all genes and the breast cancer genes is shown in Figure 5A in terms of how many different methods detect genes in these classes. We can see in Figure 5A that many genes are detected by only a few methods, whereas a small number of genes are detected by almost every method. Surprisingly, the percentage of breast cancer genes in the reported subnetworks increases with the number of methods detecting those genes, suggesting that the genes detected by more methods are more likely to be a true breast cancer genes. And also it suggests that an ensemble method that integrates multiple methods may be a better way of detecting subnetworks covering more disease genes. Similarly, we collected 2,058 interactions enriched in breast cancer pathways using KOBAS 2.0 from the KEGG pathway,7 Pathway Interaction Database (PID),16 BioCarta,17 Reactome,18 BioCyc,19 and Protein ANalysis THrough Evolutionary Relationships (PANTHER)20 databases. The distribution of interactions in terms of the number of methods detecting those interactions is shown in Figure 5B. We found that no interactions were commonly detected by more than six methods. The interactions commonly detected by more methods are slightly more likely to be enriched in pathways related to breast cancer.


Assessment of subnetwork detection methods for breast cancer.

Jiang B, Gribskov M - Cancer Inform (2014)

Number of methods detecting genes and interactions in subnetworks. Histograms of the number of genes (A) and interaction counts (B) versus the number of methods that detect them. (A) All genes denote the 7,369 genes in the HPRD network. Breast cancer genes are the 462 genes found by KOBAS in multiple disease databases. Both the gene counts are scaled to [0, 1] by dividing by the maximum count. The percentage of breast cancer genes is the breast cancer gene count divided by the count of all the genes in each category (genes found by a certain number of methods). (B) All interactions denote the 28,571 interactions in the HPRD network. Breast cancer pathways are the 2,058 interactions found by KOBAS in multiple pathways databases. Both the interaction counts are scaled to [0, 1] by dividing by the maximum count. The percentage of breast cancer pathways is the interaction count in breast cancer pathways divided by the total interaction count in each category.
© Copyright Policy - open-access
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC4256043&req=5

f5-cin-suppl.6-2014-015: Number of methods detecting genes and interactions in subnetworks. Histograms of the number of genes (A) and interaction counts (B) versus the number of methods that detect them. (A) All genes denote the 7,369 genes in the HPRD network. Breast cancer genes are the 462 genes found by KOBAS in multiple disease databases. Both the gene counts are scaled to [0, 1] by dividing by the maximum count. The percentage of breast cancer genes is the breast cancer gene count divided by the count of all the genes in each category (genes found by a certain number of methods). (B) All interactions denote the 28,571 interactions in the HPRD network. Breast cancer pathways are the 2,058 interactions found by KOBAS in multiple pathways databases. Both the interaction counts are scaled to [0, 1] by dividing by the maximum count. The percentage of breast cancer pathways is the interaction count in breast cancer pathways divided by the total interaction count in each category.
Mentions: Then we used the list of true breast cancer genes to investigate if cancer-related genes are more likely to be detected by multiple methods. The distribution of all genes and the breast cancer genes is shown in Figure 5A in terms of how many different methods detect genes in these classes. We can see in Figure 5A that many genes are detected by only a few methods, whereas a small number of genes are detected by almost every method. Surprisingly, the percentage of breast cancer genes in the reported subnetworks increases with the number of methods detecting those genes, suggesting that the genes detected by more methods are more likely to be a true breast cancer genes. And also it suggests that an ensemble method that integrates multiple methods may be a better way of detecting subnetworks covering more disease genes. Similarly, we collected 2,058 interactions enriched in breast cancer pathways using KOBAS 2.0 from the KEGG pathway,7 Pathway Interaction Database (PID),16 BioCarta,17 Reactome,18 BioCyc,19 and Protein ANalysis THrough Evolutionary Relationships (PANTHER)20 databases. The distribution of interactions in terms of the number of methods detecting those interactions is shown in Figure 5B. We found that no interactions were commonly detected by more than six methods. The interactions commonly detected by more methods are slightly more likely to be enriched in pathways related to breast cancer.

Bottom Line: Here, we compare the results of eight methods: simulated annealing-based jActiveModules, greedy search-based jActiveModules, DEGAS, BioNet, NetBox, ClustEx, OptDis, and NetWalker.While the number of genes/proteins and protein interactions detected by the eight methods vary widely, a core set of 60 genes and 50 interactions was found to be shared by the subnetworks identified by five or more of the methods.Within the core set, 12 genes were found to be known breast cancer genes.

View Article: PubMed Central - PubMed

Affiliation: Department of Biological Sciences, Purdue University, West Lafayette, IN, USA.

ABSTRACT
Subnetwork detection is often used with differential expression analysis to identify modules or pathways associated with a disease or condition. Many computational methods are available for subnetwork analysis. Here, we compare the results of eight methods: simulated annealing-based jActiveModules, greedy search-based jActiveModules, DEGAS, BioNet, NetBox, ClustEx, OptDis, and NetWalker. These methods represent distinctly different computational strategies and are among the most widely used. Each of these methods was used to analyze gene expression data consisting of paired tumor and normal samples from 50 breast cancer patients. While the number of genes/proteins and protein interactions detected by the eight methods vary widely, a core set of 60 genes and 50 interactions was found to be shared by the subnetworks identified by five or more of the methods. Within the core set, 12 genes were found to be known breast cancer genes.

No MeSH data available.


Related in: MedlinePlus