| Method |
Abstract: "[Investigators] present the first reliable estimates of the total number of phosphoproteins and phosphorylation sites (p-sites), for four eukaryotes (human, mouse, Arabidopsis, and yeast)." P.6 left column bottom paragraph: "Literature mining and stringent filtering of 187 publicly available HTP [high throughput] phosphoproteomic datasets was performed in this study so as to compile the most comprehensive data compendia for human and three model eukaryotes: mouse, Arabidopsis, and yeast. Two publicly available database compendia of low-throughput, high-quality data (from PhosphoGrid2 and Phosphosite plus), which serve as “gold-standards,” were also integrated. Based on these compendia, estimates of the total number of phosphoproteins and p-sites within each proteome were calculated using two different methods: (i) the Capture-Recapture approach that is widely used in ecology and epidemiology to estimate population size, and (ii) parameter optimization (Curve-Fitting) on the saturation curve of cumulative redundant vs cumulative non-redundant phosphoproteins/p-sites." |
| Comments |
P.3 right column bottom paragraph: "Similar analyses to those performed on the S. cerevisiae proteome were also executed with three other species. The results are presented in Figs 2 (Homo sapiens), 3 (Mus musculus), and 4
(Arabidopisis thaliana), and Table 1 compares the outcomes of the analyses of all four proteomes. In the table, the most reliable estimates, obtained by incorporating both the HTP [High Throughput] and LTP [low Throughput] non-redundant datasets, are highlighted in bold." P.5 left column top paragraph: "Notably, the maximum estimate of total p-sites significantly increased from 125,000 to 230,000 for human and from 71,000 to 156,000 for mouse. In contrast, the equivalent increase of maximum estimate
for phosphoproteins was from 10,200 to 12,800 for human and from 8,300 to 11,200 for mouse. A reasonable interpretation is that the Capture-Recapture estimates that employ the LTP data are more realistic and that the current HTP technologies alone have the potential to capture the majority of the human (80%) and mouse (74 %) phosphoproteome, but only 54% and 46% of their total p-sites. The estimates of the number of mouse phosphoproteins and p-sites are about 13% and 32% lower than
those of the human phosphoproteins and p-sites, respectively (see Table 1 for details)." See notes beneath table |