Limits...
A probabilistic generative model for GO enrichment analysis.

Lu Y, Rosenfeld R, Simon I, Nau GJ, Bar-Joseph Z - Nucleic Acids Res. (2008)

Bottom Line: This makes it hard to determine if the identified significant categories represent different functional outcomes or rather a redundant view of the same biological processes.Our model accommodates noise and errors in the selected gene set and GO.When used with microarray expression data and ChIP-chip data from yeast and human our method was able to correctly identify both general and specific enriched categories which were overlooked by other methods.

View Article: PubMed Central - PubMed

Affiliation: Computer Science Department, Carnegie Mellon University, 5000 Forbes Ave, Pittsburgh, PA, 15213 USA.

ABSTRACT
The Gene Ontology (GO) is extensively used to analyze all types of high-throughput experiments. However, researchers still face several challenges when using GO and other functional annotation databases. One problem is the large number of multiple hypotheses that are being tested for each study. In addition, categories often overlap with both direct parents/descendents and other distant categories in the hierarchical structure. This makes it hard to determine if the identified significant categories represent different functional outcomes or rather a redundant view of the same biological processes. To overcome these problems we developed a generative probabilistic model which identifies a (small) subset of categories that, together, explain the selected gene set. Our model accommodates noise and errors in the selected gene set and GO. Using controlled GO data our method correctly recovered most of the selected categories, leading to dramatic improvements over current methods for GO analysis. When used with microarray expression data and ChIP-chip data from yeast and human our method was able to correctly identify both general and specific enriched categories which were overlooked by other methods.

Show MeSH
Cell cycle comparison. Comparison of top five GO categories identified in the yeast cell cycle genes (18) by four methods. (a) Top five GO categories identified using the Classic method (hypergeometric P-value) are highlighted. Green represents the most significant category identified. The five categories represent highly redundant view of only two biological processes, as highlighted by the red circles. (b) Parent_Child method (14). (c) Weight method (15) (see website for the Elim method) and (d) GenGO. See text for discussion.
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC2553574&req=5

Figure 4: Cell cycle comparison. Comparison of top five GO categories identified in the yeast cell cycle genes (18) by four methods. (a) Top five GO categories identified using the Classic method (hypergeometric P-value) are highlighted. Green represents the most significant category identified. The five categories represent highly redundant view of only two biological processes, as highlighted by the red circles. (b) Parent_Child method (14). (c) Weight method (15) (see website for the Elim method) and (d) GenGO. See text for discussion.

Mentions: We have initially applied GenGO to analyze the well studied cell cycle expression dataset from Spellman et al. (18). We used the 800 genes determined to be cycling during the mitotic cell cycle in budding yeast. Figure 4 plots the location in the GO hierarchy of the top five categories identified by four of the five methods (see also, Table 1 and Supplementary Figure 3). The results highlight the advantages of GenGO. For example, while both GenGO and Classic successfully identify ‘mitotic cell cycle’ as the most significant category, the Classic method returns highly redundant categories including ‘mitotic cell cycle’, ‘cell cycle process’, and ‘cell cycle’. The Parent-Child method (14) also returns redundant categories (‘cell cycle process’, and ‘cell cycle’) though it does a better job in finding the more specific ‘microtubule-based process’ which is related to cytoskeleton changes during cell cycle progression (18). Both Elim and Weight fail to identify the most appropriate category for this data (cell cycle) though they do identify a number of relevant specific categories. In contrast, GenGO contains both the correct high level categories (‘cell cycle’ and ‘cell division’) as well as more specific categories (‘chromatin assembly or disassembly’) that play an important role in DNA replication and chromosome segregation. Note that cell division here is not redundant with cell cycle. While ‘cell cycle’ describes the different phases of the cell cycle, their regulation, and checkpoints, ‘cell division’ refers to the process of separation of daughter cells following the cell cycle. See Supplementary Table 3 for additional analysis of genes associated with specific cell cycle phases.Figure 4.


A probabilistic generative model for GO enrichment analysis.

Lu Y, Rosenfeld R, Simon I, Nau GJ, Bar-Joseph Z - Nucleic Acids Res. (2008)

Cell cycle comparison. Comparison of top five GO categories identified in the yeast cell cycle genes (18) by four methods. (a) Top five GO categories identified using the Classic method (hypergeometric P-value) are highlighted. Green represents the most significant category identified. The five categories represent highly redundant view of only two biological processes, as highlighted by the red circles. (b) Parent_Child method (14). (c) Weight method (15) (see website for the Elim method) and (d) GenGO. See text for discussion.
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC2553574&req=5

Figure 4: Cell cycle comparison. Comparison of top five GO categories identified in the yeast cell cycle genes (18) by four methods. (a) Top five GO categories identified using the Classic method (hypergeometric P-value) are highlighted. Green represents the most significant category identified. The five categories represent highly redundant view of only two biological processes, as highlighted by the red circles. (b) Parent_Child method (14). (c) Weight method (15) (see website for the Elim method) and (d) GenGO. See text for discussion.
Mentions: We have initially applied GenGO to analyze the well studied cell cycle expression dataset from Spellman et al. (18). We used the 800 genes determined to be cycling during the mitotic cell cycle in budding yeast. Figure 4 plots the location in the GO hierarchy of the top five categories identified by four of the five methods (see also, Table 1 and Supplementary Figure 3). The results highlight the advantages of GenGO. For example, while both GenGO and Classic successfully identify ‘mitotic cell cycle’ as the most significant category, the Classic method returns highly redundant categories including ‘mitotic cell cycle’, ‘cell cycle process’, and ‘cell cycle’. The Parent-Child method (14) also returns redundant categories (‘cell cycle process’, and ‘cell cycle’) though it does a better job in finding the more specific ‘microtubule-based process’ which is related to cytoskeleton changes during cell cycle progression (18). Both Elim and Weight fail to identify the most appropriate category for this data (cell cycle) though they do identify a number of relevant specific categories. In contrast, GenGO contains both the correct high level categories (‘cell cycle’ and ‘cell division’) as well as more specific categories (‘chromatin assembly or disassembly’) that play an important role in DNA replication and chromosome segregation. Note that cell division here is not redundant with cell cycle. While ‘cell cycle’ describes the different phases of the cell cycle, their regulation, and checkpoints, ‘cell division’ refers to the process of separation of daughter cells following the cell cycle. See Supplementary Table 3 for additional analysis of genes associated with specific cell cycle phases.Figure 4.

Bottom Line: This makes it hard to determine if the identified significant categories represent different functional outcomes or rather a redundant view of the same biological processes.Our model accommodates noise and errors in the selected gene set and GO.When used with microarray expression data and ChIP-chip data from yeast and human our method was able to correctly identify both general and specific enriched categories which were overlooked by other methods.

View Article: PubMed Central - PubMed

Affiliation: Computer Science Department, Carnegie Mellon University, 5000 Forbes Ave, Pittsburgh, PA, 15213 USA.

ABSTRACT
The Gene Ontology (GO) is extensively used to analyze all types of high-throughput experiments. However, researchers still face several challenges when using GO and other functional annotation databases. One problem is the large number of multiple hypotheses that are being tested for each study. In addition, categories often overlap with both direct parents/descendents and other distant categories in the hierarchical structure. This makes it hard to determine if the identified significant categories represent different functional outcomes or rather a redundant view of the same biological processes. To overcome these problems we developed a generative probabilistic model which identifies a (small) subset of categories that, together, explain the selected gene set. Our model accommodates noise and errors in the selected gene set and GO. Using controlled GO data our method correctly recovered most of the selected categories, leading to dramatic improvements over current methods for GO analysis. When used with microarray expression data and ChIP-chip data from yeast and human our method was able to correctly identify both general and specific enriched categories which were overlooked by other methods.

Show MeSH