By Angus Ball

Introduction

Are you surprised?? I'm not, network analysis should be interpreted even more carefully and flippantly than differential abundance. RIP. Eitherway lets get into it

Reading (and now excitingly) watching list

Barest of minimum:

This is a good paper for an oversight of what network analysis actually is (Read this paper first): Connect the dots: sketching out microbiome interactions through networking approaches

Also give this bad boy a proper read for a comparison of network analysis methods: Network analysis methods for studying microbial communities: A mini review

I've decided the package to use is NetCoMi, mostly for it's ability for comparative network analysis. If you plan on using this program go read the paper: NetCoMi: network construction and comparison for microbiome data in R

And as much as you emotionally can of the supplementary data

Shocker network analysis is hard, here are some other resources

Go watch this youtube video for a verbal explanation of network analysis: From hairballs to hypotheses: network analysis... - Karoline Faust - MICROBIOME - ISMB/ECCB 2023

Open challenges for microbial network construction and analysis

I'm so sorry... This is 19 pages of heavy hard text. I take no joy in making it a required reading and yet it is. Mercy: From diversity to complexity: Microbial networks in soils

Glossary of common terms

Network analysis is big and scary and complicated and very new to me at least. Here is a list of terms often used. This is a direct quotation from Connect the dots: sketching out microbiome interactions through networking approaches

Node - A node is a fundamental unit within the network representing an individual entity; in a microbiomea network analysis, a node is represented by a bacterial taxon at a specific level (e.g., genus, species, etc.)

Edge - In a network analysis, an edge is basically the graphical representation of an interaction between nodes, which can be positive or negative, weighted and directed or not. In a microbiome network, edges are typically just positive or negative and weighted, showing a correlation in terms of abundance.

Module - Module refers to a collection of nodes that are closely interconnected with each other, and with relatively fewer connections to nodes outside the group.

Node characteristics that define the node centrality in a network.

Degree - This refers to the count of edges that connect a chosen node with the rest of the nodes in the network. Betweenness centrality - It measures the extent to which a vertex lies on paths between other vertices/ nodes.

Closeness centrality - Reciprocal of distance sums from a specific node to all reachable nodes.

Distance – The total weight of all edges within the shortest path between two nodes.

Hub node - A node that has more connectivity within the network than other nodes, based on the node centrality values.

Keystone node - A node crucial for the observed network structure, meaning that removing this node from the network alters significatively its layout. Not all hub nodes are keystones and vice versa, as the definition of the two differs significantly. Brute force leave-one-out approaches can be used to detect this type of node

Uhm goodluck, PS hopefully you have access to a super computer (especially for the netCompare command)

NetCoMi