I'm working with bottom-up (shotgun) mass spectrometry-based proteomics and have encountered an issue. If I do a run with a HeLa digest standard, the resulting data returns many protein IDs. When I do the same experiment with cultured human cells (TCells specifically), the number of IDs drops dramatically.
We've been racking our brains trying to determine why, and it doesn't seem to be an issue with the data itself. It's as if the database search can't find sequences that fit the data.
My boss also encountered a similar issue when dealing with a previous project; the system worked well on murine samples, but there was a huge drop off in IDs when looking at human cells (with a human database, not murine of course!)
Is there some kind of known issue when looking at human cells from patients/in vitro samples? Is there a high proportion of point mutations/non-canonical sequences that upset database searches?