Help&FAQs

KnockTF constructed a TF-differentially expressed gene (DEG) network, as follows. First, for each TF knockdown/knockout dataset, we extracted DEGs under the threshold of FC≥3/2 & FC≤2/3 and formed TF-DEG pairs that were ranked based on significant levels of DEGs. Second, we combined all TF-DEG pairs for the 570 TF knockdown/knockout datasets. If a TF-DEG pair appeared multiple times in different TF knockdown/knockout datasets, we removed duplications and retained its minimum rank. Then, we reordered all nonredundant TF-DEG pairs and constructed a TF-DEG network with TFs and their DEGs as nodes and TF-DEG pairs as edges. The rank of TF-DEG pairs represented the importance of the regulatory intensity of TFs on target genes. TF-target relationships supported by the ChIP-seq data were also marked and recorded for TF-DEG pairs. Topological features such as degree, betweenness and closeness of all nodes in the TF-DEG network were computed.

4.3.1 Subnetwork Analysis

Users can submit a gene list to locate a transcriptional regulatory subnetwork. The subnetwork consists of submitted genes and their one-step neighbors within TF-DEG network. TF-target gene relationships supported by the ChIP-seq data have bold edges in the subnetwork. Users can choose subnetwork size displays by filtering the number of the most important TF-DEG pairs. KnockTF also provides topological features of subnetwork genes including degree, betweenness and closeness.

4.3.2 TF Enrichment

Users can submit a gene list and set (FDR-adjusted) P-value for TF enrichment. KnockTF maps submitted genes to the TF-DEG network and performs hypergeometric test between submitted genes and all DEGs regulated by each TF. The TFs under the threshold of (FDR-adjusted) P-value user sets are considered the most important TFs that significantly regulate the submitted genes.