Two questions about co-expression analysis

[袁鹏焜]

Dear cBioportal,

First of all, I really appreciate that your website gives lots of assistance with data analysis.

In my project, co-expression analysis (correlation) between CCL3/CCL19/CCL21/XCL1 and ITGAX/CLEC9A was performed with mRNA expression z-scores relative to all samples (log RNA Seq V2 RSEM), samples with mRNA data (RNA Seq V2) are to build patient/case set. Here are the exported results and I put them in my manuscript.

However, a reviewer gave me two questions about data analysis. I am not sure of the answers so I send this letter to seek help from you, professors.

1st:  How the correlation has been established between CCL3/CCL19/CCL21/XCL1 and ITGAX/CLEC9A  with significance via data analytics?

My preliminary idea: Give the analytical methods and algorithms based on Spearman's correlation and Pearson correlation coefficient to them, but it seems that I can not find those on your website and FAQ. It is possible for me to miss them, but could you please tell me if it is the right response to the 1st question? If it is right, could you please tell me how to explain the mechanisms of analytical methods and algorithms? 

 

2nd: To strengthen the correlation-ship, the description about TCGS data analysis component including the model parameters , is  highly desirable which is missing now.

My preliminary idea:  Tell them about my analysis process with specific parameters, such as genomic profiles and so on. I hope this is a fit response. Here are my analysis procedures.

 

Anyway, thank you for reading my letter. I study cell biology at Shanghai Jiao Tong University and I am good at doing biological experiments. I am a beginner in the field of bioinformatics data process, I hope my questions don't bother you and I look forward to your responses. Thank you very much!

 

从 Windows 版邮件发送

 

[David Higgins]

Hello!

 

1. This question was previously asked by another user, the answer is available here: https://groups.google.com/g/cbioportal/c/P-JCMNtydgM/m/1pVeo0ouBAAJ In brief, it is a 2-sided T-test, with a t-statistic derived from the correlation. Q-values are adjusted form p-values with Benjamini-Hochberg FDR correction. The full code which calculates these values can be found in our github repo: https://github.com/cBioPortal/cbioportal-frontend/blob/master/src/shared/components/plots/PlotUtils.ts#L233
2. I have not read your manuscript, so I have a hard time understanding your methods and results. I would generally say that yes, a description of how you used cBioPortal in your analysis may be suitable, including the studies you used. I would also include a description of how the Z-scores are calculated. Information about how the RNA Z-scores are calculated is included in our FAQ page, linked here: https://www.cbioportal.org/faq#what-are-mrna-and-microrna-z-scores.

 

Good luck in your research and with the publication of your manuscript!

 

David M. Higgins, Ph.D. | (he/him)
Informatics Program Manager
Center for Data-Driven Discovery in Biomedicine (D3b)
Children’s Hospital of Philadelphia, USA

--
You received this message because you are subscribed to the Google Groups "cBioPortal for Cancer Genomics Discussion Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cbioportal+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cbioportal/20220520064559.51241200B8933%40proxy02.sjtu.edu.cn.