Skip to main content
Entomology 2021 Main banner
Icon Legend
This session is not in your schedule.
This session is in your schedule. Click again to remove it.
Presentation Icons
In-Person
In-Person
Livestream
Livestream
On-Demand
On-Demand
Student Competition
Student Competition
Video Chat
Video Chat
Ticketed Event
Ticketed Event
Poster Icons
Student Competition
Student Competition
In-Person
In-Person
Poster
Poster
On-Demand
On-Demand
Q&A ViewAttendees 7

Physiology, Biochemistry, and Toxicology

10-Minute Paper

PBT: Genetics and Physiology

Ag_GCN1.0: A global gene co-expression network of the African malaria mosquito, Anopheles gambiae

Wednesday, November 3, 2021
11:00 AM – 11:12 AM MT
Location: Colorado Convention Center, Meeting Room 405-407
In-Person
Gene co-expression networks (GCNs) are used to identify potential interaction partners for genes of interest or to associate genes with unknown function to biological processes. In such networks, individual genes are represented as nodes, and significantly co-expressed pairs of genes are connected by edges.  Meta-analysis of a large number of gene expression datasets can increase the resolution of GCNs. Here, we present a global weighted GCN of Anopheles gambiae by analyzing the co-expression of gene pairs across 257 experimental conditions, published in 35 datasets. To construct the network, gene expression values were log2 transformed and z-score-normalized, which enabled comparison across different experimental designs and technology platforms. Two genes were considered significantly co-expressed when the Pearson correlation coefficient (PCC) of their expression vectors was above the fitted sliding threshold of the top 0.5th percentile of PCCs. Communities within the network were identified using the Louvain algorithm. The resulting network, which we named Ag_GCN1.0 consists of 11,825 nodes and 382,168 edges. This network is resistant to perturbation, as random removal of up to 15% of conditions lead to a reduction of only 8% of edges. In addition, we constructed 100 networks by random shuffling of expression values within conditions, which resulted in networks with less than 10 edges each, demonstrating that network topology is not due to systematic error. Ag_GCN1.0 contains 15 distinct communities, each of which is enriched significantly in unique groups of GO terms, suggesting structuring of the network by distinct biological processes.