VTC-Agg (v1.1)

SUPPLEMENTARY DATASETS

Aggregate and correlation rank grapevine gene co-expression networks.

Publication about VTC-Agg: Network aggregation improves gene function prediction of grapevine gene co-expression networks. Plant Mol Biol 103, 425–441 (2020). https://doi.org/10.1007/s11103-020-01001-2

v1.1 UPDATE: VTC-Agg is hosting an updated co-expression matrix constructed from 1,359 samples (33 experiments) using the mutual rank (MR) as the similarity metric to replace VTCdb (http://vtcdb.adelaide.edu.au/Home.aspx) which will be retiring very soon.

NETWORK CONSTRUCTION SUMMARY

Aggregate network: For each experiment (i.e. individual GCN), the Pearson’s correlation coefficient (PCC) of each gene against all other genes is first calculated and ranked according to descending PCC and subsequently thresholded to the the top 100 (stringent) and 300 (relaxed) ranked genes. This cutoff equates to a per-gene sparsity/threshold of approximately 0.34% and 1.04% of the 28,811 represented genes. For the construction of the final aggregate GCN, the frequency of co-expression interaction(s) present across individual GCNs (33 in total) were used as edge weights, ranked in descending order, and thresholded to a per-gene sparsity of 0.34% and 1.04%. See Appendix for the full list of experiment datasets used.

Mutual rank network: All 1359 samples (33 experiments) is summarized as a whole and the Pearson’s correlation coefficient (PCC) of each gene against all other genes is first calculated. The formula for MR(A,B) = √(Rank(A→B) × Rank(B→A)), was used to calculate the MR value for each pair-wise genes. Rank(A→B) corresponds to the rank assigned to gene B given the list of co-expression genes from gene A and vice versa for Rank(B→A). See above for per-gene sparsity/threshold details.

SINGLE GENE QUERY
Input the unique gene identifier for your gene (e.g. VIT_13s0067g02930) in the <Gene> box below and click the <Metric> column "AggNet" or "MR" hyperlink to retrieve the corresponding co-expressed genes lists based on aggregate network (AggNet) OR mutual rank (MR) metric. Keyword searches can also be done by through the <Functional annotation> <Network> or <BIN> boxes.

Hint: The rank of correlation measure, MR is suffice for most co-expression analysis, hence, explore this option first. The aggregate network metric, AggNet can be used for additional hypothesis generation as its adds another dimension of gene function prediction.

APPENDIX

List of microarray (29K NimbleGen Grape Whole-genome array) datasets and associated information used in the construction of the aggregate network

MICROARRAY AGGREGATE NETWORK

BULK DOWNLOAD (by chromosome)

Aggregate network thresholded at a per-gene sparsity of ~1.0% (top 300 co-expressed genes). The network file containing each gene and their relevant co-expression information is partitioned into 20 parts (Chr01 - 19, and an unplaced Chr00) according to the grapevine 12Xv1 annotation. Within each zipped file, the relevant genes and their co-expression information are organized into folders.

Chr01 (VIT_01sXXXXgXXXXX) Chr02 (VIT_02sXXXXgXXXXX) Chr03 (VIT_03sXXXXgXXXXX)
Chr04 (VIT_04sXXXXgXXXXX) Chr05 (VIT_05sXXXXgXXXXX) Chr06 (VIT_06sXXXXgXXXXX)
Chr07 (VIT_07sXXXXgXXXXX) Chr08 (VIT_08sXXXXgXXXXX) Chr09 (VIT_09sXXXXgXXXXX)
Chr10 (VIT_10sXXXXgXXXXX) Chr11 (VIT_11sXXXXgXXXXX) Chr12 (VIT_12sXXXXgXXXXX)
Chr13 (VIT_13sXXXXgXXXXX) Chr14 (VIT_14sXXXXgXXXXX) Chr15 (VIT_15sXXXXgXXXXX)
Chr16 (VIT_16sXXXXgXXXXX) Chr17 (VIT_17sXXXXgXXXXX) Chr18 (VIT_18sXXXXgXXXXX)
Chr19 (VIT_19sXXXXgXXXXX) Chr00 (VIT_00sXXXXgXXXXX)

Page updated

Google Sites

Report abuse