Table 1. Comparison of protein-protein interactions identified by
two-hybrid vs. mass spectrometry


Two-hybrid prediction

Mass spectrometry prediction



Cell cycle


Cellular process

Coenzyme A biosynthesis


Cell growth and/or maintenance

DNA repair


Response to DNA damage stimulus

DNA repair


Nuclear division

Histone acetylation


Nucleobase, nucleoside, nucleotide and nucleic acid metabolism

mRNA splicing

YDR013W (PSF1)

DNA repair

Unknown complex 1*

YJL072C (PSF2)

Unknown complex 1

Unknown complex 1*

Assignment of GO terms to uncharacterized yeast proteins based on protein-protein interactions observed by two-hybrid or mass spectrometry methods (Hazbun et al., 2003).GO Biological Process terms were determined by using the GO Term Finder available on the Saccharomyces Genome Database website (  For each protein, the tool was used to find significant shared GO terms among the interacting proteins identified by two-hybrid or mass spectrometry.  From the 100 proteins that were analyzed, 8 proteins had sufficient data that allowed the assignment of GO process terms by both methods. The protein-protein interaction overlap between the methods was low but the Biological Process terms generally agreed in that they were related and within the same branch of the ontology network.  However, the two-hybrid predictions resulted in generally lower resolution terms because each had to encompass interactions that included proteins in several related biological processes.  The asterisk indicates that these proteins were part of unknown complex 1 and could not be assigned a Biological Process term since all the complex members at the time had unknown function.