Transcript Drysdale_pheno_CSH

The FlyBase Consortium:
Harvard University
University of Bloomington-Indiana
University of California-Berkeley
Cambridge University
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005
FlyBase includes
78,804 named alleles of
19,053 Drosophilid genes
CLASSICAL ALLELES
61,286
e.g. styS73 (EMS)
GENETICALLY ENGINEERED ALLELES
17,518
e.g. styScer\UAS.cHa
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005
Core vocabularies:
Phenotypic class (fbcv.obo)
Anatomy (fly_anatomy.obo, GO)
Phenotypes are associated with
alleles using
controlled vocabularies
in conjunction with
controlled syntax
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005
The FlyBase record for styD5
(FBal0086316) includes:
Phenotypic class: neuroanatomy defective | recessive
Phenotype manifest in: chordotonal organ
Phenotype manifest in: midline glial cell
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005
Genetic interactions…for example
Casci et al., 1999 used double mutant analysis to place sty
in the EGF receptor signaling pathway (GO:0007173).
Genetic interaction: suppressor | dominant, visible { argosGMR.PC }
Genetic interaction: suppressor | dominant, eye { argosGMR.PC }
Genetic interaction: enhancer | dominant, visible { Shs.sev }
Genetic interaction: enhancer | dominant, eye { Shs.sev }
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005
Numbers …
Single mutant phenotype statements:
Phenotypic class: 86,153
Anatomy: 81,438
Total phenotypic data lines
including free text: 397,821
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005
more …
Genetic interaction statements:
Phenotypic class: 35,825
Anatomy: 39,686
Total multiple mutant phenotype lines
including free text: 131,208
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005
The Phenotype Ontology idea so far:
Observable
Anatomies
GO
Process
Function
Cell component
Cell Type
Chemical metabolite
Behavior
Temporal/developmental/life
cycle
Attribute
Value
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005
The Experiment:
“9) Now each MOD should try to
annotate 50 objects using PATO,
and swop those around …
.
.
.
Everyone do “50 units” … include
aspects of assay, etc/”
Judy Blake
May 2003 meeting’s Mins.
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005
The Questions:
1: Can the Phenotype Ontology represent all the
controlled data that FlyBase allele records include?
2: Can the Phenotype Ontology reduce the amount
of free text information in FlyBase allele records?
3: Is the proposed system efficient enough for
curators for production database use?
4: Is the end product of this type of curation
useful to biologists?
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005
The Methods:
1. Identify candidate data-rich alleles
select annotation from one reference
2. Re-annotate with Phenotype Ontology System
3. Identify
a. Structural problems?
b. Vocabulary problems?
c. Practical problems?
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005
asl1
FBal0030521
currently includes the following data, att ributed to:
FBrf0104411 == Bonaccorsi et al., 1998, J . Cell Biol.
142(3): 751--761
Phenotypic class: (with asl2) viable
Phenotypic class: (with asl3) viable
Phenotypic class: male sterile | recessive
Phenotypic class: female sterile | recessive
Phenotype manifest in: Nebenkern derivative
Phenotype manifest in: onion stage spermatid
Phenotype manifest in: onion stage spermatid &
nucleus
Phenotype manifest in: (with asl2) spermatocyte & aster
Phenotype manifest in: (with asl3) spermatocyte & aster
Phenotype manifest in: first meiotic metaphase | male
Phenotype manifest in: first meiotic anaphas e | male
“asl2/ asl1and asl3/ asl1 animals are viable.
Onion stage spermatids from homozygous males have
nuclei and nebenkern of various sizes.
Homozygous, asl 2/ asl1 and asl 3/asl1 spermatocytes are
completely devoid of asters. Homozygous
spermatocytes have morphologically normal
centrioles. However, centriole separation is abnormal;
some nebenkern are associated with two centrioles
instead of one as in wild-type. Prometaphase I,
Ontology Meeting
metaphase I and Phenotype
early
anaphase
I appear to last longer
Cold Spring Harbor
in homozygous testes
than
in wild-type."
November
19-20th,
2005
asl1
FBal0030521
Using Observable + attribute + value + qualif ier + pm ontology term:
Phenotype: (with asl2 ) adult | viability | viable | recessive
Phenotype: (with asl3 ) adult | viabilit y | viable | recessive
Phenotype: adult | femal e_fertilit y | female_sterile | recessive
OR
Phenotype: reproduction | female_fertilit y | female_sterile | recessive
Phenotype: adult | male_fertility | male_sterile | recessive
OR
Phenotype: reproduction | mal e_fertility | male_sterile | recessive
Phenotype: onion stage spermatid & Nebenkern | size | valu e\:variable??? | recessive
Phenotype: onion stage spermatid & nucleus | size | value\:variable??? | recessive
Phenotype: (with asl2) spermatocyte & aster | qu antitative | abse nt | recessive
Phenotype: (with asl3) spermatocyte & aster | qu antitative | abse nt | recessive
Phenotype: c entriole & spermat ocyte | qua litative | normal | recessive
Phenotype: c entrosome separation | process | valu e\:abnormal??? | recessive
Phenotype: m eiotic metaphase I | relative_speed | slow | recessive
Phenotype: m eiotic
Phenotype Ontology Meeting
Harbor
anaphaseCold
I | Spring
relat ive_speed
|
November 19-20th, 2005
slow | recessive
How much can we pack into one statement?
cac1 currently has
Genetic interaction:
song defective | recessive, enhanceable { mlenap-ts1 }
New system would be
Genetic interaction:
male courtship behavior (sensu Insecta)\, song production |
process | abnormal | recessive, enhanceable { mlenap-ts1 }
Point - if values were linguistically distinct then no need to state ‘attribute’ in
statement as attribute would be implicit from the attribute:value relationship.
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005
Another example
Allele combinations:
as in asl1:
Phenotype manifest in: (with asl2) spermatocyte & aster
New system would be
Phenotype:
(with asl2) spermatocyte & aster | quantitative | absent
again - better were ‘quantitative’ attribute implicit and unstated
(with asl2) spermatocyte & aster | absent
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005
Tricky to express …
Free text things:
asl1
“some nebenkern are associated with two centrioles
instead of one as in wild-type”
abnormal relationship between two observables
Bsb2
“The laterals split near the distal tip as they elongate
and at later stages they appear to be split over more
of their length”
observables showing progressive defect
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005
Caa1DAR66
“A few homozygotes exhibit neurons in the longitudinal
tracts that appear to stall at some of the commissures,
forming nodular growths”
behaviour of one observable in relation to another
‘forming nodular growth’?
egh27
“Follicle cells at the posterior of egg chamber become
mesenchymal-like”
transformation of one observable to another
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005
Equivalent alternatives:
motor_neuron ; CL:0000100
motor neuron ; FBbt:00005123
cell ; GO:0005623
cell_in_vivo ; CL:0000003
adult | male_fertility | male sterile | recessive
reproduction | male_fertility | male sterile | recessive
real issue - if redundancy exists then utility of PO for providing
comprehensive access to all phenotypes of same class is compromised
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005
Structural issue:
Sense of compound observable/attribute combination:
Observation:
“phenotype is in dorsal epithelium that makes the embryonic head”
Phenotype:
morphogenesis of embryonic epithelium | process | abnormal
Phenotype:
embryonic head epidermis | dorsal | qualitative | abnormal
but combined (better)
Phenotype:
morphogenesis of embryonic epithelium & embryonic head epidermis |
process | abnormal
issue - attribute meaningful for one but not other observable - different kind of &
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005
Curation issues:
1: Need tools
browser where only relevant values
for specific attribute are presented
“statement builder” to assemble compound
phenotype terms sequentially from vocabs.
2: Redundancy solver - would certain combinations
be disallowed in statement builder?
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005
The Answers:
1: Can the Phenotype Ontology represent all the
controlled data that FlyBase allele records include?
- potentially yes
2: Can the Phenotype Ontology reduce the amount
of free text information in FlyBase allele records?
-yes, but probably not eliminate it
3: Is the proposed system efficient enough for
curators for production database use?
- no
4: Is the end product of this type of curation
useful to biologists?
- in searches yes though redundancy needs resolution,
- for browsing? have found users to be tolerant
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005
Phenotype Ontology Meeting
Cold Spring Harbor
November 19-20th, 2005