I analyzed a group LC MS/MS data recently. After database searching with Maxquant, proteins could not been distinguished based on mass spectrum data alone were collapsed into protein groups to satisfy the principles of parsimony. I totally understand this step is necessary. However, I wondered how many proteins should I report? I read some literature and found different strategies were applied.

Some researches only reported proteins with more than two unique peptides. Some also report some with only one unique peptides after manual scrutiny. Some filtered base on grades on unique peptides (with Mascot). Some only report the so called leading proteins which have the highest number of peptides in each protein groups.

I am really confused now like a man have two watches.

What criteria should I follow to decide which proteins to report?

Any advice is appreciated.