We surveyed the "dark" proteome-that is, regions of proteins never observed by experimental structure determination and inaccessible to homology modeling. For 546,000 Swiss-Prot proteins, we found that 44-54% of the proteome in eukaryotes and viruses was dark, compared with only ∼14% in archaea and bacteria. Surprisingly, most of the dark proteome could not be accounted for by conventional explanations, such as intrinsic disorder or transmembrane regions. Nearly half of the dark proteome comprised dark proteins, in which the entire sequence lacked similarity to any known structure. Dark proteins fulfill a wide variety of functions, but a subset showed distinct and largely unexpected features, such as association with secretion, specific tissues, the endoplasmic reticulum, disulfide bonding, and proteolytic cleavage. Dark proteins also had short sequence length, low evolutionary reuse, and few known interactions with other proteins. These results suggest new research directions in structural and computational biology. structure prediction | protein disorder | transmembrane proteins | secreted proteins | unknown unknowns T he Protein Data Bank (PDB) (1) of experimentally determined macromolecular structures recently surpassed 110,000 entries-a landmark in understanding the molecular machinery of life. Structure determination lags far behind DNA sequencing, but high-throughput computational modeling (2, 3) can leverage the PDB to provide accurate structural predictions for a large fraction of protein sequences. Thus, structural data now scale with se-

No abstract

BackgroundThe majority of colorectal cancer (CRC) cases are preventable by early detection and removal of precancerous polyps. Even though CRC is the second most common internal cancer in Australia, only 30 per cent of the population considered to have risk factors participate in stool-based test screening programs. Evidence indicates a robust, blood-based, diagnostic assay would increase screening compliance. A number of potential diagnostic blood-based protein biomarkers for CRC have been reported, but all lack sensitivity or specificity for use as a stand-alone diagnostic. The aim of this study was to identify and validate a panel of protein-based biomarkers in independent cohorts that could be translated to a reliable, non-invasive blood-based screening test.Principal FindingsIn two independent cohorts (n = 145 and n = 197), we evaluated seven single biomarkers in serum of CRC patients and age/gender matched controls that showed a significant difference between controls and CRC, but individually lack the sensitivity for diagnostic application. Using logistic regression strategies, we identified a panel of three biomarkers that discriminated between controls and CRC with 73% sensitivity at 95% specificity, when applied to either of the two cohorts. This panel comprised of Insulin like growth factor binding protein 2 (IGFBP2), Dickkopf-3 (DKK3), and Pyruvate kinase M2(PKM2).ConclusionsDue to the heterogeneous nature of CRC, a single biomarker is unlikely to have sufficient sensitivity or specificity for use as a stand-alone diagnostic screening test and a panel of markers may be more effective. We have identified a 3 biomarker panel that has higher sensitivity and specificity for early stage (Stage I and -II) disease than the faecal occult blood test, raising the possibility for its use as a non-invasive blood diagnostic or screening test.

Background: Previously, we showed that gene suppression commonly occurs across chromosome 2q14.2 in colorectal cancer, through a process of long-range epigenetic silencing (LRES), involving a combination of DNA methylation and repressive histone modifications. We now investigate whether LRES also occurs in prostate cancer across this 4-Mb region and whether differential DNA methylation of 2q14.2 genes could provide a regional panel of prostate cancer biomarkers.Methods: We used highly sensitive DNA methylation headloop PCR assays that can detect 10 to 25 pg of methylated DNA with a specificity of at least 1:1,000, and chromatin immunoprecipitation assays to investigate regional epigenetic remodeling across 2q14.2 in prostate cancer, in a cohort of 195 primary prostate tumors and 90 matched normal controls.Results: Prostate cancer cells exhibit concordant deacetylation and methylation of histone H3 Lysine 9 (H3K9Ac and H3K9me2, respectively), and localized DNA hypermethylation of EN1, SCTR, and INHBB and corresponding loss of H3K27me3. EN1 and SCTR were frequently methylated (65% and 53%, respectively), whereas INHBB was less frequently methylated.Conclusions: Consistent with LRES in colorectal cancer, we found regional epigenetic remodeling across 2q14.2 in prostate cancer. Concordant methylation of EN1 and SCTR was able to differentiate cancer from normal (P < 0.0001) and improved the diagnostic specificity of GSTP1 methylation for prostate cancer detection by 26%.Impact: For the first time we show that DNA methylation of EN1 and SCTR promoters provide potential novel biomarkers for prostate cancer detection and in combination with GSTP1 methylation can add increased specificity and sensitivity to improve diagnostic potential. Cancer Epidemiol Biomarkers Prev; 20(1); 148-59. Ó2011