Hi I have a situation here which make me confused and not sure if i have every thing right.
I compared 5 different datasets in one study with each other separately. for example data1 vs data2, data1 vs data3, data1 vs data4 and data1 vs data5. Attention: All of those data were from healthy individuals which had different conditions. The purpose was to find genes which have our expected condition in healthy individuals. Then after gathering all results from above analysis we compared their readcounts with other cancer datasets(readcounts) from other studies.(All datasets in this study design were got from raw datas)
I got my result which seems already good. But I don't know in theory, my study design have any problem or not? specially the BOLD part. PCA plots in this plot shows 95% variance which make me worry.
looking forward for any suggestions from experts Kind regards.
If your data is from healthy ndividuals, why do you compare your results with cancer data set? Do you use Scanpy or Seurat to detect your DEGs? have you tried also to get DEG all over your five data sets?
By Comparing i meant getting DEG with DEseq2. for example finding gene differential expression between control(which is read count i got from those 5 condition and case which are other 15 cancer datasets. I used bowtie, featurecount and deseq2 in my analysis.
RNASeq is very prone to batch effect. It is not clear to me at all that you can distinguish between differences in your datasets cause by biology from differences caused by the samples coming from different labs. It is not clear to me that your work has much validity at all.
There is some molecular tests to prove my theory but i just want to know is it wrong to use different datasets from different labs to conducted such analysis or not.
I have no idea what question you are proposing to answer with this study. But in general, yes it is wrong to compare samples from completely different labs to each other. Again, how are you going to know which differences are biological, and which are just because the samples were prepped by different people at different places at different times?