Open Skourtis opened 1 year ago
Hi Savvas,
Good question!
You are welcome to manually remove all entries where one of the two channels is missing, this should work well for your application. If you keep them in, then as seaMass weights each peptide's contribution based on the uncertainty in quantification, it ought to adapt to these problems and downweight accordingly. So in theory it could help to keep them in, but we have not validated seaMass with SILAC data thoroughly, so cannot say for certain.
Kind regards, Andy
Hi! thanks for building this!
For SILAC quantitation, is the posterior protein intensity calculated for the heavy and liight independently and then the ratio between them is the fold change?
One of the big advantages of SILAC, is that the heavy and light ions/features are shared and traditionally only those ions which were detect in both light and heavy have and ion-level log2FC which is then summarised on the protein-level log2FC. If the protein-level posterior is done for heavy and light independently, then there is no guarantee that a posterior-light protein, had the same ions as the posterior-heavy. This might have some advantages, e.g. at very large SILAC ratios, the heavy/light ion might really not exist, giving Inf/NA ratio which is then normally discarded. I suspect that in this situation they would be included in the quant.
However, if I want to be sure that protein posterior have the same ions, could I manually remove all entries where one of the two channels (heavy/light) is missing? Would this affect the fitting of the model in a negative way?
Thank you, Savvas