GoMiner and HighThroughput GoMiner

1
GoMiner and High-Throughput GoMiner

• Paper Review
• By Chunyan Meng
• March 20, 2006

2
Paper Source

• High-Throughput GoMiner, an 'industrial-strength'
  integrative gene ontology tool for interpretation
  of multiple-microarray experiments, with
  application to studies of Common Variable Immune
  Deficiency Barry R Zeeberg 1 , and 16 others
• 1Genomics and Bioinformatics Group, Laboratory of
  Molecular Pharmacology, National Cancer
  Institute, National Institutes of Health,
  Bethesda, MD 20892, USA
• Source BMC Bioinformatics 2005, 6168
• GoMiner a resource for biological interpretation
  of genomic and proteomic dataBarry R Zeeberg1,
  and 13 others
• Source Genome Biology 2003, 4R28   

3
Outline

• Software packages overview
• Some terminology and Formula
• GoMiner Tour
• Functionality of High-Throughput GoMiner

4
Software Packages Overview GoMiner

• GoMiner helps biologicaly interpret lists of
  interesting genes (for example, under- and
  overexpressed genes from a microarray experiment)
  in the context of the Gene Ontology. It provides
  quantitative and statistical output files and two
  useful visualizations.
• The first is a tree-like structure analogous to
  that in the AmiGO browser and the second is a
  compact, dynamically interactive directed
  acyclic graph. Genes displayed in GoMiner are
  linked to major public bioinformatics resources.

5
Software Package OverviewHigh-throughput GoMiner

• High-Throughput GoMiner has capabilities
• (i) efficiently performs the computationally-inten
  sive task of automated batchprocessing of an
  arbitrary number of microarrays.
• (ii) produces a human-or computer-readable report
  that rank-orders the multiple microarray results
  according to the number of significant GO
  categories

6
High-throughput GoMiner

• (iii) integrates the multiple microarray results
  by providing organized, global clustered image
  map visualizations of the relationships of
  significant GO categories.
• (iv) provides a fast form of 'false discovery
  rate' multiple comparisons calculation.
• (v) provides annotations and visualizations for
  relating transcription factor binding sites to
  genes and GO categories.

7
Terminology and Formula
8
Terminology and Formula

• Relative enrichment factor
• Re (nf/n)/(Nf/N)
• Depletion Re lt 1
• Over-expressed enrichment (red)
• Under-expressed enrichment (green)

9
Terminology and Formula

• The two-sided Fishers exact test p-value for a
  category is used indicating the significance of
  enrichment or depletion.

10
Terminology and Formula

• False Discovery Rate q-value
• The p-value, uncorrected for multiple
  comparisons, is a measure of the statistical
  significance of a single category.
• The q-value of a category is the FDR of the list
  of categories whose p-values are equal to or
  smaller than the p-value of that category.

11
GoMiner Tour
Information visualization in GoMiner (Zeeberg et
al., Genome Biology, March 2003)
12
High-Throughput GoMiner CIM Contribution

• High-Throughput GoMiner CIM facilitates grouping
  of closely-related categories into a single
  cluster.
• High-Throughput GoMiner identifies
  biologically-relevant categories.
• The CIM also permits detection of 'cross-talk'
  between GO categories that might at first appear
  to be unrelated. For example, 'G-protein coupled
  receptor protein signaling pathway and 'cell
  adhesion' both contain the changed genes GPR56
  and CCL2.

13
Clustered Image Map
14
Transcription Factor Binding Site CIM

• The clustered image map of transcription factor
  binding sites can help to detect genomic
  regulatory networks

15
CIM Transcript Factor Binding Site
16
Time-series CIM

• High-Throughput GoMiner can be used to integrate
  time series data.

17
Time-series CIM
18
Comments

• No clear definition of FDR in the new paper and
  how that was used in the study.
• Some picture doesnt have enough explanation (the
  time-series CIM).
• Clear user interface in GoMiner.
• Multiple microarray data analysis provide more
  interpretation to the data.