search

saorisakaue / MIGWAS

A software to detect genome-wide miRNA-gene enrichment signal.
Other
8 stars 4 forks source link

Interpreting results from STEP 1/2 and paper request #2

Closed ruthchia closed 6 years ago

ruthchia commented 6 years ago

Hi, I heard this work presented at ASHG 2018 and would like to apply this to my work.

Would you be able to provide me a pre-print of your paper "Integration of genetics and miRNA-target gene network identified disease biology implicated in tissue specificity" so that I can follow the concept/idea/approach applied in this tool?

TWO additional questions:

  1. How do I interpret the values in the result files from STEP 1 in ./miRNA_P and ./gene_P? I understand that these are p-values per miRNA and per gene. If testing for only one binary phenotype, would the p value cutoff be 0.05 or do I need to adjust it for the number of genes or miRNA that was tested?
  2. And as for the results from STEP2, what is the appropriate p-value cutoff when testing only for one binary phenotype?

thanks for your time and help - looking forward to your reply.

Best wishes, Ruth chiarp@mail.nih.gov

saorisakaue commented 6 years ago

Hi Ruth, Thanks for coming to my talk and your comments!

  1. You are right, the p values from STEP1 indicate gene- and miRNA- level p values. There might not be a standard answer of the significance threshold. The pathway analysis uses the statistics as a quantitative value, rather than adopting a threshold defining whether significant or not. I did not intend to implement this single step to identify trait- associated genes/miRNAs.
  2. It also depends on your intention of the analyses. We adopted the threshold of significance at 0.05, because the analyses were permutation based and our intention was the prioritization of the important tissues associated with the phenotypes.

Best, Saori

ruthchia commented 6 years ago

Hi Saori, Thanks for taking time to explain. The output of STEP2 gives a list of tissue and the P_values and Fold_change for each tissue. How did you go from here to creating the P(MIGWAS) values you have in your paper in Figure 2a and 2b? What I am trying to get at is to understand how you aggregated the P-values for each tissue in Figure 2a and per tissue type in Figure 2b. Apologies for such basic questions – but your explanation will help me interpret your data and mine better. Thanks again Ruth

From: Saori Sakaue notifications@github.com Reply-To: saorisakaue/MIGWAS reply@reply.github.com Date: Friday, October 26, 2018 at 1:27 PM To: saorisakaue/MIGWAS MIGWAS@noreply.github.com Cc: "Chia, Ruth (NIH/NIA/IRP) [E]" ruth.chia@nih.gov, Author author@noreply.github.com Subject: Re: [saorisakaue/MIGWAS] Interpreting results from STEP 1/2 and paper request (#2)

Hi Ruth, Thanks for coming to my talk and your comments!

  1. You are right, the p values from STEP1 indicate gene- and miRNA- level p values. There might not be a standard answer of the significance threshold. The pathway analysis uses the statistics as a quantitative value, rather than adopting a threshold defining whether significant or not. I did not intend to implement this single step to identify trait- associated genes/miRNAs.
  2. It also depends on your intention of the analyses. We adopted the threshold of significance at 0.05, because the analyses were permutation based and our intention was the prioritization of the important tissued associated with the phenotypes. Best, Saori

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/saorisakaue/MIGWAS/issues/2#issuecomment-433483419, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ATxQ4SLR2qEmZNCGrFBw0gnGGLKnN56kks5uo0XRgaJpZM4X7FUn.

ruthchia commented 6 years ago

Hi Saori, Thanks for taking time to explain. The output of STEP2 gives a list of tissue and the P_values and Fold_change for each tissue. How did you go from here to creating the P(MIGWAS) values you have in your paper in Figure 2a and 2b? What I am trying to get at is to understand how you aggregated the P-values for each tissue in Figure 2a and per tissue type in Figure 2b. Apologies for such basic questions – but your explanation will help me interpret your data and mine better. Thanks again Ruth

From: Saori Sakaue notifications@github.com Reply-To: saorisakaue/MIGWAS reply@reply.github.com Date: Friday, October 26, 2018 at 1:27 PM To: saorisakaue/MIGWAS MIGWAS@noreply.github.com Cc: "Chia, Ruth (NIH/NIA/IRP) [E]" ruth.chia@nih.gov, Author author@noreply.github.com Subject: Re: [saorisakaue/MIGWAS] Interpreting results from STEP 1/2 and paper request (#2)

Hi Ruth, Thanks for coming to my talk and your comments!

  1. You are right, the p values from STEP1 indicate gene- and miRNA- level p values. There might not be a standard answer of the significance threshold. The pathway analysis uses the statistics as a quantitative value, rather than adopting a threshold defining whether significant or not. I did not intend to implement this single step to identify trait- associated genes/miRNAs.
  2. It also depends on your intention of the analyses. We adopted the threshold of significance at 0.05, because the analyses were permutation based and our intention was the prioritization of the important tissued associated with the phenotypes. Best, Saori

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/saorisakaue/MIGWAS/issues/2#issuecomment-433483419, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ATxQ4SLR2qEmZNCGrFBw0gnGGLKnN56kks5uo0XRgaJpZM4X7FUn.

saorisakaue commented 6 years ago

Hi Ruth, That is a good point. We did not aggregate the cell-type specific statistics but just annotated the cells into the tissue categories from which they are collected from. Aggregation is our future theme, but for now we did not implement this because of difference in the numbers of cells belonging to each tissue categories. Thanks, Saori