The 'ideal' dataset - Githubissues

BecksLab / ms_t_is_for_topology

🔍 A strengths and weaknesses overview of different topology generators

https://beckslab.github.io/ms_t_is_for_topology/

1 stars 0 forks source link

The 'ideal' dataset #4

Open TanyaS08 opened 6 months ago

TanyaS08 commented 6 months ago

Identifying the key features of what the 'ideal' interaction dataset would look like.

TanyaS08 commented 6 months ago

If we base this on what we need from a data input perspective it would include:

abundance/biomass (some sort of equivalent)
phylogenetic relationships (tree)
- taxonomic rank (can also be used as a 'trait'
- 'evolutionary distance'
interactions
- true positives
- true negatives (this is a bonus characteristic)
traits:
- body size/mass
- diet/feeding
- activity time
- habitat (tiering in PFIM)
- motility
- other features e.g. shells

TanyaS08 commented 5 months ago

Also need to think about 'diversity'/system we are using. Ideally I think at least one terrestrial and one marine network BUT I wonder about microbial... I think there is an argument to be had that microbial networks ≠ (going for approximate here) food webs in the traditional sense because what even is microbial feeding? But yeah....

Also maybe DINOSAUR food webs?