Is it possible to identify a protein which was revealed by western blot ?
The mass spectrometry check of the identity of a protein detected by immunoblot, which is based on a specific signal amplification technique, is really tricky. Indeed, at least ten proteins can co-migrate in the band revealed by western blot and according to their abundance they can prevent the identification of the protein of interest by mass spectrometry (dynamic range). In another words, if the protein of interest is not identify, it does not mean that this protein is not present in the band and that the antibody is not specific.
Why the molecular weight of the identified protein does not correspond to the one observed onto the gel ?
In databases, the molecular weights of proteins are calculated from the theoretical translated sequence of corresponding transcripts. All the co- and post-translational modifications of proteins, which could modify their molecular weights, are not taken into account. For example, the molecular weight of a secreted protein for which signal peptide is cleaved during its maturation or a proprotein cleaved into a mature form, is larger in the database than the one that could be detected on a gel. Conversely, a glycosylated membrane protein will have theoretical molecular weight smaller in database than the experimental one observed onto a gel.
Keratin was identified in my sample. Is this from me or you ?
To suppress the Keratin contamination we perform sample fractionation in a keratin-free room and environment. That said Keratins often come from the submitted samples, and were most likely introduced during sample preparation, or when samples are fractionated onto polyacrylamide gels during staining and cutting steps. If it is not at excess level in your samples, Keratin can be ignored. In our Sample preparation guide, we have some suggestions on how to decrease Keratin contamination on future samples.
Several proteins were identified in my gel band, how come ?
Rarely a band on a 1D gel corresponds to a single protein. Indeed mass spectrometers by far exceed the sensitivity of gel stainings (i.e., Coomassie and silver nitrate). Numerous proteins can be identified from a gel band and that depends on the complexity of the sample that was loaded onto the gel. Do not forget that the protein you are interested in may be present in amounts below the detection level of the staining method used. Overloading a gel lane is also a classical error that will affect the resolution and separation of proteins. Generally, it also results in carry over of abundant peptides in Liquid Chromatography ahead of the mass spectrometer, resulting in the repeated identification of the same protein in different samples.
How can I be sure you correctly identified my protein ?
Protein identification is performed by seeking in databases theoretical peptide spectra corresponding the best to acquired experimental peptide spectra. This spectra comparison is performed with dedicated softwares, which give scores reflecting the probability for the matches to be right. Moreover wrong identifications were estimated by calculating a false discovery rate in the dataset (FDR). Identifications are then validated with best matches and a FDR < 1%. The identification of proteins with these criteria and several peptides is confident. For proteins identified with only one peptide, in general it should be necessary to check manually the spectra annotation.
Is it possible to sequence a protein by mass spectrometry ?
Mass spectrometry classical proteins analyses are not strictly speaking sequencing methods. These analyses are based on experimental and theoretical spectra matching for protein identification, as explained in the previous question. We only obtained in rare conditions (pure sample in sufficient amount and proteoform listed in database) a sequence coverage of 100% with a classical proteomics analysis allowing to be sure of the sequence of the proteoforms detected.
Is it possible to identify proteins from non-sequenced organisms ?
Protein identification is based on database searching, it is thus not possible in principle to identify proteins with no information listed in database. Even a query in a known neighboring species are in general unfruitful because spectra comparison is based on mass matching and the substitution of only one amino acid change the mass of the peptide, thus preventing its identification. We recommend for this kind of organisms to have some transcripts information at least (RNAseq).