I have a list of refseq numbers each with a KPRM value before and after a transformation of cells in an attempt to make immoralised cells. I need to work out which of these genes have changed expression and are transformed, but i am unsure how to do this as i havent been given any numerical information about how different the expression it needs to be for it to count as 'changed'. I have tried using logs and ploting various graphs but i need to get some more exact info.
You need a lot more information to perform significance testing for differential expression:
How many replicates do you have? By the sounds of it, you only have one replicate for each condition (before/after transformation), which means that you don't know the variability of your observations. This makes it impossible to ascertain significance; large differences may be spurious if your variability is high.
What values do you have? I've never heard of KPRM, I assume you mean RPKM if you're working from RNA-seq data. For significance testing with packages like edgeR and DESeq, it is necessary to have the original count data so that the mean-variance relationship can be properly modelled.
More generally, I suggest you read the user guides for packages that are commonly used for detecting differential expression. I've already mentioned edgeR and DESeq(2), but there's also limma (made suitable for RNA-seq data via voom). These will define a set of differentially expressed genes based on some significance cut-off, most typically by applying a 5% threshold on the false discovery rate.
