SMD : Help : Experiment Set Help

Help : Experiment Set Help

Contents

Description
Data Organization
Documenting the Experimental Factors of the Set
Experiment Set Annotation and Description
Experiment Set Display and Analysis
Editing Experiment Sets
Deleting Experiment Sets
Common problems

Related Help Documents

Convert an Arraylist to an Experiment Set: A brief description of the whys and wherefores of converting an arraylist or result set list into an experiment set.

Description

Using experiment sets, you can save a group of arrays you access frequently when viewing or clustering data. In additional, experimental sets are used to group array data within the context of a publication. This help page describes how to create and use experiment sets.

Sets can be used for several purposes:

They can be used to organize experiments more flexibly than either just category or subcategory (the only restriction is that the arrays be from the same organism).
They can serve a similar purpose as ArrayLists, but can be shared with the group/public without need of a local UNIX account.
They facilitate a more accurate documentation of the experimental design for a set of hybridizations.
They can help assist in data organization prior to publication.

Data Organization

Using the Advanced Results Search, select your desired search parameters and click "Data Retrieval and Analysis". On the subsequent Search Results page, select all experiments (from one organism, only) you want to organize and click "Create Experiment Set". The selected experiments are placed in the right box on the Experiment Set Organization page. The order within the 'Experiments Included within Set' box, from top (first) to bottom (last) is the order that experiments will be displayed and clustered (if experiment re-ordering is not specified during clustering). Using the various buttons you can:

Order all your selections alphabetically either ascending (Sort Asc) or descending (Sort Desc)
Move selected individual or groups of experiments up or down, ordering within the 'Experiments Included within Set' box.
Remove unwanted experiments (or hold them for later re-addition) within the 'Removed Experiments' box using the 'Remove (All)' and 'Add (All)' buttons.

After achieving the desired order of arrays within the 'Experiments Included within Set' box, click one of the following submission buttons.

Select Experimental Factors
Selecting experimental factors is recommended for those sets that are destined for publication, or indeed any set where the experimenter wishes to fully document the experimental design, including experimental factors, factor values, and measurements.
Save Experiment Set
By clicking the 'Save Experiment Set' button, the experimental factor selection step is omitted, and the user skips to the final step, Experiment Set Annotation

Experimental Factor Selection and Description

Experimental factors can be selected after organization of the experiment set (prior to the final set annotation), or by editing an existing experimental set. Experimental factors are the dependent and independent variables of the experiment itself (Ex. time, glucose concentration, disease state, etc.) Each time experimental factors are associated with an experiment set, three may be designated. If more than three are required to fully explain the experimental design, then the newly created set must be edited, and additional factors added as required.

Selecting Experimental Factors and Descriptions:
1. First, choose the experimental procedure or design that contains the experimental factors which the set is organized from the first column's menu.
2. Next choose the experimental parameters from the second column's menu. This will be the experimental factor, whose value is often systematically varied among the arrays contained within the set. If the experimenter had previously annotated the individual array's experimental parameters, these parameter/factor values will be pre-filled within the final annotation step. For more on pre-association of an array's procedural details, please consult the helpfile on entering procedural annotation or edit the details of each array (via Display Data).
3. If the factor chosen is not sufficiently descriptive of the experimental design, or if it requires either qualification or contextual explanation, supply a description of the experimental factor within the textbox immediately below to the factor menu.
Measurements: If your chosen experimental factor has a measurement (i.e. it will likely have a value with associated units for all arrays), then select the unit kind, default unit, and unit type from the pop-up menus in the far-right column.
- If your factor is not measured, make certain the top units-menu in that row is set to "Not measured".
- The default unit selected can be altered on a per array basis on the next step. Do not fret if the unit doesn't apply to all arrays within the set. Unit kinds (time, concentration) and types (absolute, relative), however, do apply to all arrays.
- If the appropriate unit kind or default unit is not available within the pop-up menus, please suggest to the database curators that your unit of choice be included.

Experiment Set Annotation and Description

Once the experiment order has been established for the set, and the experimental factors have been selected (optional), now the appropriate annotations need to be made.

First and foremost, an appropriate Experiment Set Name (required) must be entered. Currently, the only method of returning experiment sets is by name, through the Basic Search. Therefore, it is important that a succinct, self-explanatory name is chosen (try not to be arbitrary or ambiguous). Please use these guidelines in order to optimize the query/display within the Basic Search:
1. Try to limit the name to 60 characters (max 100).
2. The organism does not need to be included within the Experiment Set Name.
3. Keep it succinct (use the Description field (below) for your elaborate details).
Longevity: As there is a need for temporary worksets of data, you can choose to make the set permanent, or self-removing (after a couple weeks).
Experiment Set Design (required) - Experiment Sets are required to be classified as a specific type, in order allow better display of the contents.
1. Replicate : Biological - This set type is for repetition of the experiment prior to extracting mRNA (typically) for analysis. These sets are meant to assay the reproducibility in biological experimental conditions.
2. Replicate : Technical - This type is for an repetition of experimental conditions after mRNA extraction. This includes probe labeling, hybridization, and reverse duplicates.
3. Logical Grouping - This type describes to experiments that "logically" go together when viewed within a larger context (similar to Categories encompassing subcategories).
4. Time Course - A set of arrays where the primary experimental factor is Time (temporal measurements). Often includes an additional experimental factor, such as Compound Based Treatment, Growth Conditions, or Developmental Stage.
5. Dose Response - A set of arrays where the experimental factor is a concentration measurement variable. Often includes a Compound Based Treatment experimental factor.
Clustering Weight : If the weight of an experiment should be reduced (if it is represented more than once within the set) or increased, enter the factor by which its weight should be multiplied. For example, entering "0.5" would halve the weight an experiment has on clustering. The default weight is 1.
Factor Values : For each factor that was selected, and for each array, the user can specify an optional value. The factor value can be a text string or numeric. In the case of a measurement it is intended that the value be a number only, as units are found in the adjacent pulldown menu (if the factor was specified as being measured on the previous step). For example, for the array representing the 10 minute timepoint within a Time Series set, enter 10 for that factor's value and select m from the adjacent unit menu.
Description : This field provides for the full, exhaustive description that the experiment set name does not allow. This field will be viewable within 'View Experiment Set' from the list display. Think of this field as a figure legend or an abstract.

Experiment Set Display and Analysis

Once the experiment set has been created, the experiments contained can be displayed and clustered within the set's context. Currently, the only method of returning experiment sets is through the Basic Search. To either display or analyze experiment sets, within the "Pick results type" section at the top of the Basic Search page select one of the following:

My Experiment Sets : limits the browsable select lists to the sets you created. Hence, this option is limited to internal researchers.
Experiment Sets : lists all viewable experiment sets, by organism, within the browsable select lists.

As in the advanced search, there are two choices to retrieve the data:

Display Data (table list): This search submission will return the experiment contents of the experiment set, as well as optional links to view and edit/delete the experiment set details (creator only).
Data Retrieval and Analysis: This Basic Search submission is the launch point to begin data analysis. All the experiments contained within the set are passed the directly to data selection prior to clustering, as it is assumed the user wants the analysis to be conducted within the context of the set's organization. For further information on subsequent steps in clustering, see the Advanced Results Search help.

For further information on the Basic Search, see the Basic Search help.

Editing Experiment Sets

Only the creator of the experiment set has the ability to add or remove experiments, re-arrange the experiment order, or edit the set's annotation. To edit a experiment set you own, use the Basic Search, choose the experiment set name from the select list, and Display Data (as described above). At the top of the resulting display page, a header containing the set's Name, Organism, and Type are displayed, along with its Options. If you are the set's owner, there is an

icon. Clicking on this icon will open a pop-up window containing 4 editing options whose behavior is very similar to the process of experiment set creation, detailed above:

Edit Experiment Set Annotation : use this option to edit the name, experimental design, factor descriptions, factor values, units, and description of the set (without changing the experimental contents or order).
Rearrange/Remove experiments from experiment set : use this option to either remove unwanted experiments or rearrange the order of experiments within the experiment set. In addition, after you do this, you are allowed to either add/edit experimental factors or edit the set annotation, if desired.
Add/Remove Experimental Factors : choose this option to either document additional experimental factors or delete prior factors that were poorly conceived or annotated. Note: deletion of experimental factors will remove all previously associated factor values for all arrays if the changes are commited upon the final annotation step.
Add new experiments into experiment set : use this option to add new experiments to the existing set from the Advanced Search (limited to the set-specified organism). After choosing the desired experiments and clicking "Edit Experiment Set", the experiment additions are in the left box while the pre-existing experiments in the pre-specified order are on the right. Integrate the new experiments into the set by clicking the "Add All" button and use the move buttons to re-order the set, if needed. Saving the new set will pass you to Option 1 above, allowing you to edit the set annotation, if desired.

The behavior of these three commands is very similar to the process of experiment set creation, detailed above. As in set creation, you may not introduce experiments from different organisms into the set using "Add new experiments into experiment set".

Deleting Experiment Sets

Only the creator of the experiment set has the ability to delete the experiment set. To delete a experiment set you own, use the Basic Search, choose the experiment set name from the select list, and Display Data (as described above). At the top of the resulting display page, the experiment set header contains the Options. If you are the set's owner, there is an

icon. Clicking on this icon will remove the experiment set organization and annotation, but will leave the experiments intact.

Common Problems

Currently experiment sets may only include experiments from a single organism.
Both the interactive ordering/movement of arrays within Experiment Set Organization and the heirarchical menus of procedures/parameters and measurement-kinds/units with experimental factor selectionrequires JavaScript be turned on in the web browser.
- For Netscape Navigator/Communicator:
  1. On the Edit menu, select Preferences.
  2. Under Web Browser, click Web Content.
  3. In the Active Content area, select the 'Enable scripting' check box.
- For MS Internet Explorer:
  1. On the Edit menu, select Preferences.
  2. Click Advanced.
  3. Select 'Enable JavaScript'

There are required fields within Experiment Set Creation. You must specify the Experiment Set name, and the Classification. If you specify your saved set as temporary, it will be deleted after two weeks.

Please send comments or questions to: array@genome.stanford.edu