Skip to content

Latest commit

 

History

History
138 lines (87 loc) · 6.42 KB

session4_DE_module_in_Foundry.md

File metadata and controls

138 lines (87 loc) · 6.42 KB
Raw

Running the DE Module in Foundry

Before you start

Please go to https://viafoundry.umassmed.edu/ and login into your account. If you have a login issue, please let us know about it ([email protected]). We will set up an account for you.

Creating a Run

Once logged in, click on the Projects section at the top menu and either 1. select an existing project or 2. click the Add a New Project button and enter your project name and click OK. To access pipelines, click the Pipelines tab and then click the Add Pipeline button.

Search for DE module and click on the Add button on "DE module" and close the window.

Now click the Run button of the DE module pipeline on the table.

  1. The run page will be loaded. Give your run a name
  2. Under Run Environment, select "New UMASS SCI Cluster"
  3. Enter your work directory on the cluster. (e.g. /home/{your_cluster_username}/foundry/)

Count Matrix

DESeq2 requires a count matrix as input. In this file, there is one column per sample and one row per feature as well as a header line that contains the sample names. Each cell contains the counts for a particular feature for each sample. The first four lines look like this:

gene transcript D01_Ctrl_0h D01_Lps_1h D01_Lps_2h D01_Lps_4h D01_Lps_6h D09_Ctrl_0h D09_Lps_1h D09_Lps_2h D09_Lps_4h D09_Lps_6h
C9orf152 NA 0 0 3 0 5 1 0 6 6 2
RPS11 NA 11825 13716 7204 4339 11187 5855 6940 4941 7354 6446
ELMO2 NA 773 805 372 490 1776 518 615 347 949 1190

Note: The second column ("transcript") must exist for this module. This column is included to be consistent with the output of some feature counting programs which store information here. It will be removed by the module during processing.

To add the counts matrix, do the following:

  1. Under User Inputs, next to Counts file, click Enter File
  2. Click the green Add File button in the top right to enter new files.
  3. Next to "1. File Location", enter:
/share/data/umw_biocore/class/de_module_demo/
  1. and click the magnifying glass button. The box below should populate with files like so:

  1. Next to 2. File Type, choose TXT
  2. Next to 3. Collection Type, choose Single/List

  1. Under 4. File Pattern select donnard_counts.txt and click "Add Selected Files"

  1. Under 5. Collection Name give the collection the name donnard_counts
  2. Click "Save Files"

  1. Click "Save"

Group File

This module requires some information about each sample in order to group the samples together by condition. At minimum it must contain one column named sample_name and one column named group. All samples listed in this file must have a corresponding column in the count matrix. The group column is used to notate which samples belong to the same condition (replicates). Additional metadata columns are allowed. An example looks like this:

sample_name group batch condition time
D01_Ctrl_0h ctrl A Ctrl 0h
D01_Lps_1h 1h A Lps 1h
D01_Lps_2h 2h A Lps 2h
D01_Lps_4h 4h A Lps 4h
D01_Lps_6h 6h A Lps 6h
D09_Ctrl_0h ctrl B Ctrl 0h
D09_Lps_1h 1h B Lps 1h
D09_Lps_2h 2h B Lps 2h
D09_Lps_4h 4h B Lps 4h
D09_Lps_6h 6h B Lps 6h

To add the grouping file, do the following:

  1. Next to Groups file click Enter File and add the following file:
/share/data/umw_biocore/class/de_module_demo/donnard_metadata.txt

Comparison File

The module needs to know which groups to compare. This file must contain columns named controls, treats, and names. Each row represents a comparison of the treatment group to the control group. The values of controls and treats must exist in the group column from the group file. The names column sets the prefix for output files. An example looks like this:

controls treats names
ctrl 1h 1h_v_ctrl
ctrl 2h 2h_v_ctrl
ctrl 4h 4h_v_ctrl
ctrl 6h 6h_v_ctrl

To add the comparison file, do the following:

  1. Next to Comparison file click Enter File and add the following file:
/share/data/umw_biocore/class/de_module_demo/donnard_comparisons.txt

Genome Build

This example was performed in humans. In the Genome build dropdown select human_hg38_gencode_v34.

DESeq Settings

There are several parameters that affect how the module will behave. This example experiment includes a time course of LPS treatment in duplicate. The complete time course was performed on two separate days (one replicate from each day). Batch correction may be necessary with this design. To override the default parameters for DESeq2 (enable batch correction):

  1. Click the wrench next to next to Run DESeq2

  1. Navigate to the Batch correction options section and enable batch correction.
  2. In the PCA plot options, set "Color" to batch and "Shape" to group.

Run Pipeline

The run is now ready. Click Run in the top right corner. The DE module typically takes 2-3 minutes to complete for this dataset.

Viewing Results

When the run is complete, navigate to the results tab and look at the report generated for each comparison.

Congratulations! You have run and tested the DE module pipeline on Foundry!