GENE-E may be downloaded and used free of charge by
academic and other
non-profit researchers. When referencing your use of GENE-E, please cite this website.
Import Data
Use the File menu to open a data file. See the file formats page for supported formats. Example data files can be
loaded by selecting File>Open Example Data.
Annotate Columns/Rows
Annotate columns or rows based on entries provided in a tab delimited text file or an Excel .xls or .xlsx
file. A colored bar visually identifies members of the same category.
Categories are primarily used for visualization. In analyses such as RIGER, column categories can also be
used to identify phenotypes.
To annotate columns/rows:
- Create a tab delimited text file or an Excel .xls or .xlsx file.
- Select File>Annotate Columns or File>Annotate Rows and open the file created previously.
GENE-E displays color bars below the column names or to the right of the row names to indicate the categories
to which the columns/rows belong.
Select Edit>Column Annotations or Edit>Row Annotations to edit the color for a category or delete a
category.
GENE-E displays column annotations below the column names and row annotations to the right of the row names
New Heat Map
To open a new heat map on a subset of your data:
1. Select the desired columns and rows.
2. Select Tools>New Heat Map.
Preferences
Select GENE-E>Preferences (Mac) or View>Options (other platforms) to modify the title, look and feel of the
current visualization tool. GENE-E displays a window which provides options specific to the current visualization tool.
Most options are self explanatory.
The color tab controls the colors used in the heat map:
Relative: GENE-E converts values to heat map colors using the mean and maximum values for
each row or the standard deviations from the row mean for each row (as determined by the settings on this tab).
Global. GENE-E converts values to heat map colors using the minimum and maximum values in
the entire data set (as determined by the settings on this tab).
To change the color of the heat map
click a colored square above the heat map legend and select a new color. Click and drag a
colored square to move a control point. Click the add button to add a new control point. Click delete to delete an
existing color.
Sorting
You can sort columns by column name, category, or annotation. You can sort rows by row name, category, annotation, or
the values in a particular column.
To sort columns:
Select Tools>Sort Columns.
Select the field(s) to sort by. Each drop-down list includes column (for column name) and all categories and
annotations that you have loaded (in this example, the Phenotype category).
To sort rows:
Select Tools>Sort Rows or click on a row header (shift-click to add a secondary sort).
Select the field(s) to sort by. Each drop-down list includes row (for row name), each column name, and all
categories and annotations that you have loaded.
Mask Data
Masking rows or columns temporarily hides them from many GENE-E operations. For example, masked rows and columns can be
omitted from new heat maps (Tools>New Heat Map).
Highlight one or more rows or columns.
Select Tools>Mask Rows or Tools>Mask Columns. Alternatively, right-click and select Mask Rows or Mask
Columns from the context menu.
You can clear masked columns/rows by highlighting one or more columns/rows and selecting Tools>Unmask
Columns or Tools>Unmask Rows or to clear all masked columns/rows select Tools>Clear Column Mask or Tools>Clear
Row Mask.
RIGER
RIGER ranks shRNAs according to their differential effects between two classes of samples, then identifies the genes
targeted by the shRNAs at the top of the list. In this way, RIGER identifies genes essential to the difference between
the classes. For details, see Luo, Cheung, Subramanian, et al. (2008).
- The Class Comparisons parameter defines one or more class comparisons on which to run RIGER. Click Edit to
open the Edit Comparison window. You can select columns by categories or by name. For each class comparison, you select
columns for the two classes to be compared: Highlight the desired categories (or columns) in the middle panel. Click an
arrow icon to move them to class A or B. To remove categories (or columns) from a class, highlight them and click the
delete icon. GENE-E moves them back to the middle panel.
- Optionally add another two class comparison by clicking the add button.
- Select a method to analyze your RNAi screening data.
Second Best Rank
A method based on ranking genes by the rank of the second best scoring hairpin for that gene. This is
currently the preferred method for RNAi screening analysis in the RNAi Platform.
Kolmogorov-Smirnov
A statistical method and the basis for the original RIGER as described in Luo et al, PNAS, 2008
Weighted Sum
This method is a modification of the Second Best Rank in that it takes the combined sum of the first and
second best ranks for hairpins for a given gene. The best ranking hairpin is given a weight of 0.25 and the second best
ranking hairpin is given a weight of 0.75. The sum of these weighted ranks is used to compute a new score, and genes
are ranked by this new score.
Regardless of the method you choose to use, you may use either Signal to Noise or Log Fold Change to generate
a single hairpin level score from the set of replicates for each hairpin in your screen. All the methods, including
Kolmogorov-Smirnov, can use either Signal to Noise or Log Fold Change. You may also use the T-Test option, although
this is still experimental and has not been fully tested.
Please note that we now compute the FDR as a p-value, and therefore additional permutations are required for
the null distribution.
When it is finished, RIGER results are listed in the Navigator window as shown below. Double-click the 'NES'
or 'Scored shRNAs' to view the results.
Scored shRNAs
Double-click 'Scored shRNAs' to display a heat map showing the ranked list of shRNAs and their scores. The
heat map is sorted by the scores of the first comparison. Select Tools>Sort Rows to sort a different comparison by
score.
NES: Double-click 'NES' to display a heat map showing the normalized enrichment scores (NES) for the genes in
your dataset. The heat map is sorted by the NES of the first comparison. Select Tools>Sort Rows to sort a different
comparison by the NES.
Analysis Report
To view a report of RIGER results, right-click either RIGER result node and select View Gene Report or View
Hairpin Report. To save the report, right-click either RIGER result node and select Save Gene Report or Save Hairpin
Report. The report lists: Gene, Hairpins, # Hairpins, # Hairpins in the top 500, 1000, 5000, and 10000, NES, Gene rank,
p-value, and p-value rank.
Marker Selection
Marker selection identifies entities that are differentially expressed between two classes. For each entity, the
analysis uses a test statistic to calculate the difference in expression between the classes and then estimates the
significance (p-value) of the test score. It then corrects for multiple hypotheses testing (MHT) by computing both the
false discovery rate (FDR) and the family-wise error rate (FWER).
The output of marker selection consists of:
Score: The calculated value of the test statistic.
p-value: The estimated significance of the test statistic for this row (not yet corrected
for MHT).
p-value low: The estimated lower bound for the p-value.
p-value high: The estimated upper bound for the p-value.
FDR(BH): The expected proportion of non-marker genes (false positives) within the set of
genes declared to be differentially expressed. It is estimated using the Benjamini and Hochberg procedure. (Benjamini,
Y. and Hochberg, Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing.
Journal of the Royal Statistical Society. Series B (Methodological). 57(1): p. 289-300.1. 1995.)
FWER: The probability of having any false positives.
Hierarchical Clustering
Hierarchical clustering recursively merges objects based on their pair-wise distance. Objects closest together are
merged first, objects furthest apart are merged last. The result is a tree structure, referred to as a dendogram, where
the leaf nodes represent the original items and internal (higher) nodes represent the merges that occurred.
Boxplot