federicomarini/pcaExplorer

Multifactorial exploration -

jamesthomas232 opened this issue · 11 comments

jamesthomas232 commented

image

Hi guys,
When trying to carry out a multifactorial exploration, I cannot select factors to compare, and I when I type them they disappear.
Any help would be appreciated.

federicomarini commented

Hi @jamesthomas232,
I cannot see too much from the screenshot, nor you did provide enough info to get a grasp of the dataset at hand.
Can you elaborate more on this maybe?
Federico

  • 👍1
jamesthomas232 commented

Hi @federicomarini,
Thank for getting back so quickly! I've sent a new screenshot, hopefully it will be more clear!

image

I have 19 samples (10 disease, 9 control) and ~56000 probes. Cell type refers to disease or no disease (annotated as IPF/Con) and Gender refers to Male or Female (annotated M/F)
I was aiming to use the multifactorial expression hopefully with these factors.

Is there any other information that could help? Please let me know!

jamesthomas232 commented

And to clarify, I can neither select or type the factor available levels, and if I try to compose the matrix, it says "Error: a dimension is zero"

federicomarini commented

Thanks for the larger screenshot, this way I can have a cleaner look at it.

Hm, strange, it looks like it is not able to pick the values of the levels for the factor?

When I run it on the example dataset (from the airway package) I see something like this:

image

Can you reproduce that? (Load the demo, compute the vst data first)
If not, the problem can be more general. If you can get there, it might be something specific to the dataset you have at hand somehow.

HTH,
Federico

jamesthomas232 commented

I can reproduce that with the demo package, I can select the factor available levels too, which is what I am struggling to do with my own dataset.
Many thanks,
James

federicomarini commented

Hm, ok, this is somehow tricky then.
Can you check that the factors are encoded as such (if you are providing the data as dds/transformed)?

Or otherwise: can you provide me with a smaller/anonymized version of the dataset you are using? If so, please reach out via email (you can see it in the maintainer field of the package description)

mars188 commented

Hello, I am facing the same issue as above. Were you guys able to solve this?

When I try to select factors, it doesn't allow on my data, however, it works fine with example airway data.

Please let me know! Many thanks!

mars188 commented

I had my raw counts data and metadata in .csv format but then I saved it as Text (tab delimited) format and uploaded to pcaExplorer web.

Then I faced the same issue as mentioned above.

jamesthomas232 commented
federicomarini commented

I am afraid I cannot do too much without checking directly how the dataset seems to be formatted.

As I said above: a smaller/anonymized version of the dataset might be the way to go for me to try and help you a little more?

mars188 commented

Dear @federicomarini

I have sent you a smaller/anonymized version of the dataset on the given email address (marinif@uni-mainz.de). Please take a look at it and let me know the issue.

This multifactorial analysis is something crucial for me and pcaExplorer seems to provide the solution to what I am looking for.

Many thanks for your time and help.