Email updates

Keep up to date with the latest news and content from BMC Systems Biology and BioMed Central.

Open Access Highly Accessed Research article

Global organization of protein complexome in the yeast Saccharomyces cerevisiae

Sang Hoon Lee1, Pan-Jun Kim2 and Hawoong Jeong3*

Author Affiliations

1 IceLab, Department of Physics, Umeå University, 901 87 Umeå, Sweden

2 Institute for Genomic Biology, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USA

3 Institute for the BioCentury and Department of Physics, Korea Advanced Institute of Science and Technology, Daejeon 305-701, Korea

For all author emails, please log on.

BMC Systems Biology 2011, 5:126  doi:10.1186/1752-0509-5-126

Published: 15 August 2011

Additional files

Additional file 1:

Strength distributions of (a) complex-mode projection and (b) proteins-mode projection (Figure S1). Here, the strength in (a) corresponds to the sum of number of proteins shared with the neighboring complexes for each complex, and the strength (b) corresponds to the sum of number of complexes shared with the neighboring proteins for each protein. The blue squares correspond to the cumulative strength distribution P(s) = ∑s′ s p(s′), and the pink lines and gray curves are the best exponential and power-law fittings, respectively.

Format: EPS Size: 51KB Download file

Open Data

Additional file 2:

Projected network's degree distribution from the generating function approach (Figure S2). Assuming that both protein and complex's degree distributions follow the exponential degree distribution p(k) ~ exp(-k), the projected network's degree distribution is numerically calculated with the generating function approach mentioned in the main text. One can clearly observe that the degree distribution follows the exponential tail for k ≫ 1.

Format: EPS Size: 21KB Download file

Open Data

Additional file 3:

Degree distribution of complexes in the bipartite network for E. coli protein complex data in (a) semi-log scale and (b) double-log scale (Figure S3). Here, the degree corresponds to the number of component proteins for each complex. The blue squares correspond to the cumulative degree distribution P(k) = ∑k′ k p(k′).

Format: EPS Size: 23KB Download file

Open Data

Additional file 4:

Degree distribution of complexes and proteins in the bipartite network for human protein complex data, in comparison with the protein-protein interaction network (Figure S4). Here, a complex's degree (blue square) corresponds to the number of component proteins for each complex, and a protein's degree (red circle) corresponds to the number of complexes in which a protein participates as a component. The degree distribution of proteins in the binary protein-protein interaction is shown as pink triangles, as a comparison.

Format: EPS Size: 22KB Download file

Open Data

Additional file 5:

List of all the protein complexes with MIPS functions assigned by our method (Table S1). List of all the protein complexes with MIPS functions assigned by our method, where the core and attachment components are taken from Ref. [16] (and the indices are the same as Ref. [16]). We classify each function into the following three categories. (1) Gavin_1st (light yellow): minimizing the number of complexes for each newly assigned function, described as Eq. (4) in the main text. (2) Gavin_2nd (light green): minimizing the number of proteins for each newly assigned function, instead of that of complexes (3) The functions assigned by both (1) and (2) (bright yellow). Note that we only select high confidence (HC) outcomes among the raw outcomes, whose reliability of function assignment considering the multiple solutions is large.

Format: XLS Size: 244KB Download file

This file can be viewed with: Microsoft Excel Viewer

Open Data

Additional file 6:

List of all the proteins with MIPS functions newly assigned by our method (Table S2). List of all the proteins with MIPS functions newly assigned by our method (bright yellow), along with the ones from the MIPS database (sky blue) which is used as "input function." HC outcomes are selected as in Additional file 5, Table S1. Sometimes the functions already annotated in MIPS and the ones assigned by our method are quite similar, and very different in some cases. Therefore, we suggest the latter case be worth investigating further, as we did in the main text.

Format: XLS Size: 616KB Download file

This file can be viewed with: Microsoft Excel Viewer

Open Data

Additional file 7:

Statistical validation of abundance estimation (Figure S5). For each value of fraction of training set p, the relative deviation

is ranked and shown compared to its random counterparts (meaning that the identity of proteins with estimated abundance is randomly paired with the ones with real abundance values). The real deviation values are always on the left side (smaller than) of the random counterparts outside the error range, which implies its statistical significance.

Format: EPS Size: 76KB Download file

Open Data

Additional file 8:

Statistical validation of the function assignment, in comparison with the hypergeometric test in Ref. [23](CYC2008) (Table S3). Here p is the fraction of training set. Fraction of test proteins with at least one function assigned, which has at least one assigned function on the original MIPS functional datasets. *Among all the newly assigned functions to the entire test set proteins, fraction of functions which are also on the original MIPS functional datasets. Raw outcomes correspond to the original outcomes from our method, and high confidence (HC) outcomes are the selected subset of them whose function assignments are invariant even if the multiple solutions are considered.

Format: XLS Size: 26KB Download file

This file can be viewed with: Microsoft Excel Viewer

Open Data