Comparing RNA expression between different tissue type, but same Cancer.

Question

[9521ljh]

i have 40 SARCOMA data. and there are many groups of tissue.

if i use RNA expression normalization method, is it okay to compare?

More specific, if i have 12 bone tissue sarcoma and 28 Uterus tissue type.

using these 40 sample, the edgeR results in MDM4 gene is differentially expression in tumor than normal.

i think it is nonsense to compare.

am i right?

Answer 1

[i.sudbery]

Sarcoma just means a cancer of non-connective tissue. So bone sarcoma and uterine sarcoma are not the same cancer. However.... If you have matched tumour and normal for each sample (that is, you have 12 bone normals and 28 Uterus normals), there is no statistical reason you can't compare both tissue types together. Note however, that as you have 28 uterine and only 12 bone, lots of your results would probably be uterine genes, whose effect sizes had been diluted by the presence of the bone samples.

You could have tissue type as an explanatory variable. You would probably then find genes that are common cancer genes (i.e. likely upregulated in any cancer).

[9521ljh]

Sincerely really thank you!