Hello to all biostars,
For months I have managed to do this with expression patterns and now I have to interpret it, but it generates doubts and I really want to interpret it in the best way, or be as accurate as possible.
My data are from stages of cell development, and what we are looking for is the relationship between the different stages and the unique patterns that these genes show, of course we have 13 but while I deduce it I wanted to know a good suggestion or things that I should take previously when doing it. And what we are looking for is based on epithelial architecture.
Now a little context to know how these patterns were achieved: 11 MDCK cell samples were sequenced and identified with a program that describes their morphology by many physical parameters such as height, circularity, etc. These cells were grouped into 4 groups, each with 2 replicates except for the 2nd plot, which has 4. Subsequently, a filtering of 10% variability in gene expression was done, basically making an average, then the standard deviation was divided with the average. . After that, the clusering was done, and here there was a debate for weeks with my boss because I suggested normalizing the sample with log2 or deseq2, but doubts arose that log2 is a transformation (which it is), but people describe it as normalization , so in the end we use scaling. Then we graph the samples by clustering to then identify the plot that you now see. Now all those genes were put in Gprofile and since it was difficult to identify them all, revigo was used which makes the biological processes that occur a little more understandable (it is worth mentioning that here we only use BP)
