Weighted analysis
0
0
Entering edit mode
14 months ago
Peter • 0

I am a newer bioinformatician so my knowledge is still growing.

We have bulk rna seq data from two datasets and are hoping to compare them. Is there an analysis method that weights each DEG in regards to LFC and pvalue? When we previously compared the two dataset we just looked at the DEGs which ignored two columns of data specifically LFC and pvalue.

Any help is appreciated. Thank you!!

RNA-seq • 827 views
ADD COMMENT
2
Entering edit mode

This sounds like an XY problem. (https://xyproblem.info/). What are the two datasets (two technologies, same samples, you want to compare whether one is better; two separate experiments with the same contrast, you want to find "reproducible" signals; two related phenotype vs control contrasts, you want to find similarities and differences)?

ADD REPLY
0
Entering edit mode

Sorry about that. First time posting and didn't really know how to phrase things.

We have two bulk rna datasets that come from liver samples. One is a cancer model and the other is a liver regeneration model. We are hoping to compare the two to see similarities between the dataset to see if there is a correlation between cancer progression and liver regeneration. When we compare the DEG we are ignoring the weight of significance for each DEG (ie lfc, pvalue, padj) and just using a venn diagram to plot making it just a two dimensional analysis. Is there a way to incorporate lfc, pvalue, padj when comparing two dataset and how would i plot to visualize it?

ADD REPLY
1
Entering edit mode

When we compare the DEG we are ignoring the weight of significance for each DEG (ie lfc, pvalue, padj) and just using a venn diagram

Do you mean that, instead of using "Is GeneX differentially expressed Y/N?" instead using the extent of differential expression (i.e., effect size)?

The only effect size is the log-fold change; the p-values are a measure purely of confidence.

Have you tried, among significant genes, looking at the log-fold change correlations? Have you plotted the log fold change values against one another, and tried to segment genes into concordant (same direction of DE) and discordant genes? You can also evaluate ranks by asking whether the MOST DE genes in liver regeneration correspond to the MOST DE genes in cancer or not.

Just some ideas.

ADD REPLY
0
Entering edit mode

Thank you. I appreciate the help.

ADD REPLY

Login before adding your answer.

Traffic: 1326 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6