Deion of the bug Hello, I'm running this pipeline on the

<a class="user-mention notranslate" data-hovercard-type="user" data-hovercard-url="/us

"exclude" samples aren't respected in SHINYNGS_STATICEXPLORATORY - Error in cor(matrices[[g]]) : supply both 'x' and 'y' or a matrix-like 'x' about differentialabundance HOT 6 CLOSED

olgabot commented on September 27, 2024

"exclude" samples aren't respected in SHINYNGS_STATICEXPLORATORY - Error in cor(matrices[[g]]) : supply both 'x' and 'y' or a matrix-like 'x'

from differentialabundance.

Comments (6)

pinin4fjords commented on September 27, 2024 1

Thanks for the report and the detective work- I'll look into this soon!

from differentialabundance.

olgabot commented on September 27, 2024

Hello, I've done a bit of digging, and I think this error is happening because exclude samples in my --contrasts file aren't getting excluded in the ShinyNGS Exploratory Plot generation.

Digging into one of the runs, I saw that this plot has three categories: green, blue, and red. But it should only show blue and red, the green samples shouldn't be there in this comparison:

My contrasts file (redacted) is like this, as a nicely formatted table:

id	variable	reference	target	blocking	exclude_samples_col	exclude_samples_values
A_vs_B	A_vs_B	groupA	groupB		A_vs_B_exclude	TRUE
B_vs_C	B_vs_C	groupB	groupC		B_vs_C_exclude	TRUE

And then the samplesheet is like this:

sample	fastq_1	fastq_2	A_vs_B	A_vs_B_exclude	B_vs_C	B_vs_C_exclude
groupA_sample1	groupA_sample1_R1_001.fastq.gz	groupA_sample1_R1_001.fastq.gz_R2_001.fastq.gz	groupA	FALSE		TRUE
groupA_sample2	groupA_sample2_R1_001.fastq.gz	groupA_sample2_R1_001.fastq.gz_R2_001.fastq.gz	groupA	FALSE		TRUE
groupA_sample3	groupA_sample3_R1_001.fastq.gz	groupA_sample3_R1_001.fastq.gz_R2_001.fastq.gz	groupA	FALSE		TRUE
groupB_sample1	groupB_sample1_R1_001.fastq.gz	groupB_sample1_R1_001.fastq.gz_R2_001.fastq.gz	groupB	FALSE	groupB	FALSE
groupB_sample2	groupB_sample2_R1_001.fastq.gz	groupB_sample2_R1_001.fastq.gz_R2_001.fastq.gz	groupB	FALSE	groupB	FALSE
groupB_sample3	groupB_sample3_R1_001.fastq.gz	groupB_sample3_R1_001.fastq.gz_R2_001.fastq.gz	groupB	FALSE	groupB	FALSE
groupC_sample1	groupC_sample1_R1_001.fastq.gz	groupC_sample1_R1_001.fastq.gz_R2_001.fastq.gz		TRUE	groupC	FALSE
groupC_sample2	groupC_sample2_R1_001.fastq.gz	groupC_sample2_R1_001.fastq.gz_R2_001.fastq.gz		TRUE	groupC	FALSE
groupC_sample3	groupC_sample3_R1_001.fastq.gz	groupC_sample3_R1_001.fastq.gz_R2_001.fastq.gz		TRUE	groupC	FALSE

So I think the issue is in providing matrices for EXPLORATORY_PLOTS from SHINYNGS, to remove the exclude samples. Or to provide a non-null value for the exclude samples -- I'll try this for now and let you know.

from differentialabundance.

olgabot commented on September 27, 2024

Update: Still getting the same error, even when the empty/null/None/NA values in A_vs_B are replaced with "Other":

sample	fastq_1	fastq_2	A_vs_B	A_vs_B_exclude	B_vs_C	B_vs_C_exclude
groupA_sample1	groupA_sample1_R1_001.fastq.gz	groupA_sample1_R1_001.fastq.gz_R2_001.fastq.gz	groupA	FALSE	Other	TRUE
groupA_sample2	groupA_sample2_R1_001.fastq.gz	groupA_sample2_R1_001.fastq.gz_R2_001.fastq.gz	groupA	FALSE	Other	TRUE
groupA_sample3	groupA_sample3_R1_001.fastq.gz	groupA_sample3_R1_001.fastq.gz_R2_001.fastq.gz	groupA	FALSE	Other	TRUE
groupB_sample1	groupB_sample1_R1_001.fastq.gz	groupB_sample1_R1_001.fastq.gz_R2_001.fastq.gz	groupB	FALSE	groupB	FALSE
groupB_sample2	groupB_sample2_R1_001.fastq.gz	groupB_sample2_R1_001.fastq.gz_R2_001.fastq.gz	groupB	FALSE	groupB	FALSE
groupB_sample3	groupB_sample3_R1_001.fastq.gz	groupB_sample3_R1_001.fastq.gz_R2_001.fastq.gz	groupB	FALSE	groupB	FALSE
groupC_sample1	groupC_sample1_R1_001.fastq.gz	groupC_sample1_R1_001.fastq.gz_R2_001.fastq.gz	Other	TRUE	groupC	FALSE
groupC_sample2	groupC_sample2_R1_001.fastq.gz	groupC_sample2_R1_001.fastq.gz_R2_001.fastq.gz	Other	TRUE	groupC	FALSE
groupC_sample3	groupC_sample3_R1_001.fastq.gz	groupC_sample3_R1_001.fastq.gz_R2_001.fastq.gz	Other	TRUE	groupC	FALSE

from differentialabundance.

pinin4fjords commented on September 27, 2024

So, I can't reproduce this error exactly. For example, I unset the sample grouping variable in the test profile for one sample, and whilst it plotted a group with an empty legend entry, I didn't get an error at the dendrogram stage.

I'm making a PR here to prevent unlabelled samples getting plotted, and I notice a fix is also needed for GSEA.

It's important to note that the exploratory script isn't designed to plot per contrast, just per unique contrast variable, and does not pay any attention to the exclude_samples column in the contrast.

I hope that the above PR may help your case (once we update the module, and update the module in the workflow), but if not you will need to provide me with something reproducible for me to do more.

from differentialabundance.

pinin4fjords commented on September 27, 2024

@olgabot could you test with current dev to see if the issue persists after the above fixes?

from differentialabundance.

pinin4fjords commented on September 27, 2024

Closing for lack of activity

from differentialabundance.

"exclude" samples aren't respected in SHINYNGS_STATICEXPLORATORY - Error in cor(matrices[[g]]) : supply both 'x' and 'y' or a matrix-like 'x' about differentialabundance HOT 6 CLOSED

Comments (6)

Related Issues (20)

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent