SEARCH

SEARCH BY CITATION

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

FilenameFormatSizeDescription
pmic7305-sup-0001-FigureS1.tif1031KFigure S1. One iteration of determining one cluster. It shows four steps: (a) Project models to the centroid and calculate the first Gaussian fitting of inline image. (b) Choose the initial model set Dset1, which includes models with Dscore1 in the range of inline image. (c) Choose the improved model set Dset2, which includes models with small Dscore1 (less than σ) to inline image. The centroid of inline image is treated as the reference Dref, i.e., the estimated center of the biggest cluster. (D) Construct the cluster by collecting models close to Dref, i.e., with Dscore1 to Dref less than 2σ.
pmic7305-sup-0002-FigureS2.tif1644KFigure S2. Dscore distribution and comparison. (a) Comparison between RMSD and Dscore1(with a Pearson correlation coefficient of 0.9338), with two linear fitting lines and the density plot. A grid of 0.1 Å is used to generate the density. (b) Comparison between TM-score and Dscore2 (with a Pearson correlation coefficient of 0.9220), with a linear fitting line and the density plot. A 0.05 interval of TM-score is used to generate the density.
pmic7305-sup-0003-TableS1.doc32KTable S1. The one closest to the centroid vs. medoid for I-TASSER model clusters
pmic7305-sup-0004-TableS2.doc32KTable S2. Pearson/Spearman correlation coefficients between Dscore1 and RMSD, and between Dscore2 and TM-score
pmic7305-sup-0005-TableS3.doc106KTable S3. Comparison between MUFOLD-CL and SPICKER in CPU time, RMSD and TM-scores of the representative of the largest cluster
pmic7305-sup-0006-TableS4.doc60KTable S4. Comparison between MUFOLD-CL and SPICKER in RMSD and TM-scores of the best representative of top-3 largest cluster for 49 proteins with 20000 or less models
pmic7305-sup-0007-TableS5.doc34KTable S5. Performance comparison among MUFOLD-CL, SPICKER, Pleiades and Calibur on nine hard proteins
pmic7305-sup-0008-TableS6.doc39KTable S6. Comparison between MUFOLD-CL and ONION in CPU time, cluster number, and TM-scores of the representative of the largest cluster
pmic7305-sup-0009-TableS7.doc66KTable S7. RMSD and TM-score of the selected representatives using Dscore1 and Dscore2
pmic7305-sup-0010-TableS8.doc56KTable S8. MUFOLD-CL clustering results for two Rosetta model sets, with 500,000 models each for protein 1aoy_ and 1abv_
pmic7305-sup-0011-S1.doc34KSupplimentry Material

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.