#12 SAINTexpress Imputed values
I'm not sure I understand why the imputed values in SAINTexpress are the average of the prey intensities with missing values as this doesn't seem to agree with the original method paper https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3744231/. It seems to give a very high value that produces some strange results. For example these two results:
553F PROT1 PROT1 0.092|0.379|0.367 0.839 0.280 3 .|.|. 0.000 0.000 0.000 0.000 0.000 -14.838 0.580 0.891
553F PROT2 PROT2 0.385|0.525|. 1.392 0.464 3 .|.|. 0.000 0.000 0.000 0.000 0.000 -14.838 0.962 0.886
Both have measurements in the tagged experiments and no measurements in the control but as SAINTexpress chose the imputed value as 0.482224 for all missing data the resultant fold changes somehow come out less than 1 and they are obviously called insignificant.
Maybe I am missing something but should the imputed values not be at the lower end of the detection ability of the machine rather than the average of measurements that could be missing at random.
Thanks
