Summarize Inference Report
The Summarize Inference Report module allows the user to summarize the results of an Inference Report, by counting and reporting the number of rows that meet a set of criteria (i.e. p-value, fold change, estimate, etc.). The result is an interactive table containing the number of rows that meet each specified criteria, for each of the selected estimates in the inference report. Table results can be used to generate lists for quick filtering.
To run this module, type MicroArray | Inference | Summarize Inference Report .
Input Data Requirements
This module works on Inference Reports.
- Project & Data: The window includes a drop-down box to select the Project and Data object to be filtered.
- Rows: Selections can be made on which variables should be included in the Summarize Inference Report (options include all, selected, visible, and any pre-generated Lists).
- Estimates: The user must also choose the Estimates (tests) from the Inference Report that should be included in the summarization process. Any available estimates from the selected report will be shown in this box.
- In the Inference Table, users can set filters based on combinations of different criteria. Potential criteria include Estimate, Fold Change, Raw P-value, Adjusted P-Value, and Max(LSMean). For instance, if the user wanted to count the number of rows with a Fold Change >2 and a Raw P-value <.05, they should check both the Fold Change and Raw P-value boxes. Then, enter nothing in the Fold Change< box, and 2 in the Fold Change > box. Finally, the user should enter 0.05 in the Raw P-value< box. Clicking Add will add this condition to the Conditions section, where the user has the choice (recommended) of naming each condition.
- The Max(LSMean)> box will only have an effect under the following conditions: 1. The user ran an LSMeans test while running the GLM or One-Way/Two-Way ANOVA modules; 2. The user chose to append LSMean data to the inference report. Max(LSMean) > will count the number of rows where, for each estimate, the maximum LSMean value for all the LSMean values in that estimate is greater than a specified value.
- Add: Add current criteria into conditions.
- Remove: Remove selected criteria from conditions.
- Clear: Clear all criteria from conditions.
- Load: load a previously saved "Conditions" text file.
- Save: save the currently specified conditions to a text file to a local directory.
- For each condition, report a common list that are present in at least _ estimates: This option allows the user to report a combined list if a particular gene appears in a number of the estimates (as defined by the user).
- Generate lists based on conditions: Checking this box will automatically create one List per condition/per estimate. Note: This has the potential to generate an enormous number of lists, and thus is not selected by default. Also, lists can be manually generated once the summarized table is created, as this table is interactive with Details On Demand. The user can Select the list folder to place the generated lists in a specifically organized folder.
- Select list folder: users can specify the folder to generate the lists.
An example of the output is:
Example:Identify all significant probesets
In an inference report with tests for each timepoint, the user may want to find the genes that are significantly up-regulated in at least one timepoint.
In the Summarize Inference Report window, select the Estimates (tests) to query, and specify the cutoff parameters (e.g. at least 4-fold up-regulated AND the difference has a p-value < .05), and click Add:
In the Advanced tab, the user can optionally select For each condition, report a common list that are present in at least 1 estimates, which will generate a List item containing all genes passing the criteria in at least one test.
Alternatively, the user can select all rows of the resulting Summary Table, which will display the union of the significant genes from each test, in the Details window:
The user can generate a list from selected rows to create a new list of genes, which can be used to filter any table or -Omic data for the significant gene set: