[IGBF-3916] Molly's R script on Brandon's nextflow results - JIRA UNCC

Details

Type: Task
Status: Closed (View Workflow)
Priority: Major
Resolution: Done
Affects Version/s: None
Fix Version/s: None
Labels:
None

Story Points:
3
Epic Link:
Support NSF pollen grant
Sprint:
Fall 2

Description

Goal:
To run the Deseq2 script that Molly perfected on the next flow results of the reference free assemblies. The goal here is get Brandon more acquainted with R and to see what difference we can observe within a variety.

e.g., Heinz versus Heinz de novo

To start, Brandon has a copy of Molly's script, and the tomato data for the Muday lab. the salmon counts table and the metafile needed to run the script.
1. Get this to run as intended.
2. Make a metafile for the Heinz versus Heinz de novo run.
3. Tweak the Heinz versus Heinz de novo nextflow table to be suitable for this script.
4. Run the script!

Attachments

Activity

Ascending order - Click to sort in descending order

Robert Reid created issue - 20/Sep/24 3:09 PM

Robert Reid made changes - 20/Sep/24 3:09 PM

Field	Original Value	New Value
Epic Link		IGBF-2993 [ 21429 ]

Robert Reid made changes - 24/Sep/24 8:58 AM

Status

To-Do [ 10305 ]

In Progress [ 3 ]

Hide

Permalink

Robert Reid added a comment - 03/Oct/24 9:50 AM

Script was pulled down via the github tools built into Posit's Rstudio software.

The original code ran smoothly producing the expected results.

Trying the code on the de novo assembly data, Brandon ran into a few issues related to the headers of the columns.
We resolved this via a few carefully placed greps within the R code (thanks Steven).

At quick glance, de novo read counds and ref based read counts are entirely different due to the process each dataset goes through. These are not really comparable via EdgeR or deseq2.
So new plan to highlight bias has been devised and this task is no longer important.
Case closed.

Show

Robert Reid added a comment - 03/Oct/24 9:50 AM Script was pulled down via the github tools built into Posit's Rstudio software. The original code ran smoothly producing the expected results. Trying the code on the de novo assembly data, Brandon ran into a few issues related to the headers of the columns. We resolved this via a few carefully placed greps within the R code (thanks Steven). At quick glance, de novo read counds and ref based read counts are entirely different due to the process each dataset goes through. These are not really comparable via EdgeR or deseq2. So new plan to highlight bias has been devised and this task is no longer important. Case closed.