Behavioral phenotypes & patterns

[matentzn]

(copied from some old meeting minutes I have now retired)

• Initial focus of behavior terms should be movement behaviors - coming from big screens

• We will need behavior terms in order to define these new phenotypes.

• We need to avoid getting too hung up on where these behavior terms live. There are 3 possibilities:

  • A new fly behavior onotology
  • Extend GO
  • Extend NBO
  • -> Agreement on Phenotype call was to prefer GO if possible but use NBO for more species specific. Wherever there is an GO term, NBO should use it. We may need to work with GO on alignment. We will soon start up a regular meeting NBO/GO and behavior ontologies in general

Some Relevant branches:

• NBO kinesthetic behavior
• NBO:body part movement
• GO:locomotory behavior

Relevant tickets

• add behaviour phenotype terms for FlyBowl dataset
• Document larval locomotor modes for defining phenotypes #10
• Headless fly behaviour - activation screen #22

Behavior phenotype patterns

• Lets try building a DOSDP pattern for some simple kind of behaviour

Screens generally measure an increase or decrease in some movement, e.g.

Q: Increased frequency of 
E:  wing flicking
Qual: Abnormal

Sometimes we need social/environmental context:

Near wall
male behavior during pursuit of female
Male behavior while facing female 

• Fold into behavior terms or post-compose?
• What relations to use?

FlyBowl

• Unsupervised learning revealed behaviour that people have not noticed before

• Winflick Glossary

• Behaviour Map

  • Thousand driver lines that drive expression in one of that lines, and than you check what kind of behaviour emerges
  • They have expression patterns in terms, they can correlate
  • Regions of expression patterns where you are most likely to get wing flick

• Headless flies

  • Goal: What circuits in the ventral nervous system drive the movement
  • Manually curated

• Should we create our own fly behaviour ontology

  • Is NBO aligned with GO? To some extent, but needs work
  • Wherever there is a GO behaviour, it should be in NBO

• UBERON-CARO tension

  • Granularity divergence
  • Reduce ontological commitment in higher level terms
  • Storage of memory vs recall
  • GO term should encompass both?

How to fold in anatomy:

• Decision: Avoid Uberon references in NBO for now.
• Use FBbt to fold in classification by anatomy - e.g. wing flicking under wing behavior phenotype vis FBbt

@Clare72

• Try to group the behaviours in some kind of sensible way

Increased Frewuency of Kinethetic movement
…
…
Priority on things that we have circuits on
Olfactory behaviour is weird.

• Unified semantics of sex modelling