some doubts using the platform

[Luís Martins]

Hi,

I have been exploring this tool to analyse my proteomics data and would like to ask a few questions to understand how some calculations are performed and how some data is presented to better interpret my data. 

1 - In a contrast where both the control and main groups have more than one condition, how are the conditions in the same group computed?  Is there a statistical calculation that accounts the weight of the conditions of that group or is it simply an average of all the samples in that group independently of their condition?

 

2 - Regarding the WGCNA and the modules feature, after computing to include 8000 genes I obtain 18 different modules but only 12 are shown in the Module-Trait relationship chart. I would like to ask what is the selection criteria for the modules that are shown?

 

3 - Also in the WGCNA - Modules - Enrichment plot, only for module 1 I see some of the top most enriched genesets bar in pink, for all other modules all geneset bars are shown in blue. I would like to ask what is the meaning of this difference in colour?

 

Thank you in advance for your help.

 

Kind regards,

 

Luis Martins

[BigOmics Analytics Team]

Hi Luis,

Mind you, the WGCNA is still beta. So we are still working on it.

For question 1. Can you please give a concrete example? I am not sure what you mean with this.

For question 2 & 3, it would be helpful to show screenshots. 

Best

Ivo

[Luís Martins]

Hi Ivo,

Thanks for your reply. 

For question 1:  I have 2 diseases that I want to compare and each of the diseases has variants. For this I created a contrast where the control group has 2 variants ( belonging to disease 1

) and the main group has 3 variants (belonging disease 2). You can find the example in the attached figure 1. My doubt is how omics playground computes the variants in each group? Does it just average everything in that group or is there a specific statistical test applied?

Please find attached some screenshots representing my questions.

Thank you.

Best wishes

Luis

[BigOmics Analytics Team]

About your first question, if you create a control/main group using multiple variants it will just pool all samples that match as one group. There is no sample or subgroup weighting. Thus if the variants are unbalanced then the mean is biased towards the bigger variant subgroup. So again, it is simply pooling all samples together.

For the other questions, please be patient, I need to look into the code. If you want, and are apt, you can probably figure out what's happening. I guess, I have defined a max number of modules to avoid overcrowding of the heatmap

[Luís Martins]

Thanks again for your help. I will wait patiently for your input on how the beta features work. 

Best regards,

Luis