Following the report of Mycobacterium tuberculosis proteins found in archaeological bone by Boros-Major et al. (2011), we attempted to identify M. tuberculosis proteins in mummified lung tissues from which ancient DNA success had already been reported. Using a filter-aided sample preparation protocol modified for ancient samples we applied shotgun proteomics to seven samples of mummified lung, chest and pleura tissues. However, we only identified four peptides with unique matches to the M. tuberculosis complex, none of which were unique to M. tuberculosis, although we did identify a range of human proteins and non-mycobacterial bacterial proteins. In light of these results, we question the validity of the peptide mass fingerprint (PMF) approach presented by Boros-Major et al. (2011), especially because the PMF spectrum presented in Boros-Major et al. (2011) has similarities to that of human collagen, the dominant protein in the tissue under investigation. We explore the challenges of using proteomic approaches to detect M. tuberculosis, and propose that, given the contentious outcomes that have plagued ancient protein research in the past, the susceptibility of ancient material to modern contamination, and the degradation inherent in archaeological samples, caution is needed in the acquisition, analysis and reporting of proteomic data from such material.