monarch-initiative / monarch-disease-ontology-retired Goto Github PK

From @mellybelly on May 27, 2015 14:46

http://www.rarechromo.org/html/DisorderGuides.asp
(may also have additional phenotypes for curation)

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#36

From @nlwashington on March 6, 2015 17:34

Here's the KEGG disease id to OMIM disease id map:
http://rest.genome.jp/link/disease/omim

note that is is a 1:many relationship (so they are often grouping classes).

there are some distinct disease-gene associations in kegg, so having their disease ids would be really helpful.

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#3

From @mellybelly on March 14, 2015 21:53

For example:
OMIM_107730 Apolipoprotein B; APOB
subclass of DO:disease and Orphanet_121386 (no label in file, but is a gene from Orphanet apolipoprotein B)

coming from MGI file I think, genes need to be pruned out upstream.

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#14

From @cmungall on March 18, 2015 15:38

Just noticed our pipeline was hardwired to download v1.2...

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#21

From @cmungall on March 13, 2015 6:22

We should manually curate groupings at the level above DC

See for example, subtypes of epilepsy, that should be under DOID:1826 ! epilepsy syndrome

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#11

for @mellybelly

From @cmungall on March 17, 2015 21:47

• DC is probably the same as the ORDO here
• the ORDO hierarchy doesn't join up as well as it should here. Not much we can do about this.

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#20

From @cmungall on July 20, 2015 22:32

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#38

From @nlwashington on March 6, 2015 17:25

we can use the phenotypic series from OMIM as some grouping classes. The phenoseries ids are part of the omim file, here:
http://nif-crawler.neuinfo.org/monarch/ttl/omim.ttl

see ticket here:
obophenotype/obsolete-hdo#25

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#1

From @cmungall on April 20, 2015 20:23

Attached is the hierarchy for OMIM:236670-Muscular Dystrophy-Dystroglycanopathy

Yellow is ORDO. Note that there is too much of a lattice here. We have glycosylation multiple places as siblings.

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#30

From @cmungall on May 16, 2015 0:57

We are missing some links here;

• blue = DO
• yellow = ORDO
• brown = OMIM
• grey = MESH

Should DOID:9182 ! pemphigus be merged into Orphanet:79669 ! Autoimmune bullous skin disease?

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#34

From @cmungall on July 20, 2015 22:53

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#39

From @cmungall on March 9, 2015 1:7

Consider the phenotype parathyroidism and its occurrence in various diseases:
http://www.monarchinitiative.org/phenotype/HP:0000843
(click 'disease' tab)

These are the types of DC:0000201 'hyperparathyroidism'

• OMIM:145000 ! HYPERPARATHYROIDISM 1
• OMIM:145001 ! HYPERPARATHYROIDISM 2
• OMIM:239200 ! Hyperparathyroidism, Neonatal Severe Primary
• OMIM:600166 ! Hyperparathyroidism, Primary, Caused By Water Clear Cell Hyperplasia
• OMIM:610071 HYPERPARATHYROIDISM 3

One might think we can use 'hallmark' to cluster diseases. However, the hallmark qualifier is lacking from phenotype association for the phenotype hyperparathyroidism for the above diseases.

For comparison, Orphanet classifications of OMIMs:

• OMIM:145000 ! HYPERPARATHYROIDISM 1 Orphanet:99877 ! Familial parathyroid adenoma,Orphanet:99879 ! Familial isolated hyperparathyroidism
• OMIM:145001 ! HYPERPARATHYROIDISM 2 Orphanet:99880 ! Hyperparathyroidism - jaw tumor syndrome
• OMIM:239199 ! Hyperparathyroidism, Neonatal Self-Limited Primary, with Hypercalciuria NOT-IN-ORDO
• OMIM:239200 ! Hyperparathyroidism, Neonatal Severe Primary Orphanet:417 ! Neonatal severe primary hyperparathyroidism
• OMIM:256120 ! Nephropathy, Deafness, and Hyperparathyroidism Orphanet:2668 ! Nephropathy - deafness - hyperparathyroidism
• OMIM:600166 ! Hyperparathyroidism, Primary, Caused By Water Clear Cell Hyperplasia Orphanet:99878 ! Primary parathyroids hyperplasia
• OMIM:612089 ! Hypophosphatemic Rickets And Hyperparathyroidism NOT-IN-ORDO

To complicate matters, it seems we have no annotations for OMIM:610071 HYPERPARATHYROIDISM 3 (hence it doesn't show up above)

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#8

From @cmungall on March 27, 2015 13:27

Compare

[Term]
id: OMIM:604403
name: Generalized Epilepsy With Febrile Seizures Plus, Type 2
is_a: DC:0000159 ! Febrile Seizures, Familial
is_a: DC:0000160 ! Epilepsy, Generalized, with Febrile Seizures Plus

With

[Term]
id: OMIM:604233
name: Generalized Epilepsy With Febrile Seizures Plus, Type 1
is_a: DC:0000160 ! Epilepsy, Generalized, with Febrile Seizures Plus

Is this intentional, or is this just lack of consistency over how restrictive the parent (DC:0000159 ) is?

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#24

From @cmungall on April 7, 2015 2:47

There are many HP and DOIDs with the same label (thankfully HP uses a leading upper case, usually annoying, but this saves us in Protege, where the same label causes massive issues).

I would rather see explicit different labels otherwise this will cause massive confusion.

Also these equivalent labels could be used to generate axioms - not equivalent classes, but something like 'has hallmark phenotype'?

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#29

From @cmungall on March 17, 2015 18:42

How is this coming along?

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#19

From @nlwashington on March 6, 2015 17:27

Add mappings between ORDO and DO.

see obophenotype/obsolete-hdo#15

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#2

From @mellybelly on March 31, 2015 17:10

As part of of the build we need to pull the definitions from DO, OMIM, and orphanet, with some order of priority. @nlwashington has built the OMIM and GeneReview text descriptions, these should be relatively easy to pull together with DO terms? then we could potentially identify terms with no definitions (eg. DC terms).

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#25

From @mellybelly on March 13, 2015 1:6

OMIM URIs are currently mostly leaf nodes (or merged as equiv classes if name is same as parent and there is only one child). However, there will occasionally be terms in DC or DO with the same label that don't meet requirements for the merge. Since we don't want to make equivalencies for these OMIM terms (or Decipher for that matter), these terms should be appended with (OMIM) in the label to distinguish them from their parents.
For example, you could have:

DOID_1932 Angelman's syndrome
--OMIM_105830 Angelman's syndrome
--Orphanet_98794 Angelman syndrome due to maternal 15q11q13 deletion

Here for example, we might like to have OMIM appended like this:
DOID_1932 Angelman's syndrome
--OMIM_105830 Angelman's syndrome (OMIM)
--Orphanet_98794 Angelman syndrome due to maternal 15q11q13 deletion

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#9

From @nlwashington on March 6, 2015 17:50

i've been looking through our the ontology, i've found some issues.

many of the classes in the merged.obo are missing seemingly pedantic grouping classes.

for example,

id: OMIM:614480
name: Hypertriglyceridemia, Transient Infantile
is_a: DOID:630 ! genetic disease

id: OMIM:145750
name: familial hypertriglyceridemia
def: "A lipid metabolism disorder characterized by elevated triglyceride levels as a result of excess hepatic production of VLDL or heterozygous LPL deficiency." [url:http://en.wikipedia.org/wiki/Familial_hypertriglyceridemia]
comment: OMIM mapping confirmed by DO. [SN].
synonym: "Pure hyperglyceridemia" EXACT []
xref: DOID:0050527
is_a: DOID:3146 ! lipid metabolism disorder
is_a: DOID:630 ! genetic disease

are not related to one another except for both being genetic diseases. why isn't the former also a lipid metabolism disorder? or even better, why aren't they both classified as "hypertriglyceridemia"? that seems like a natural grouping class.

another

id: OMIM:606170
name: Genitopatellar Syndrome
is_a: DOID:630 ! genetic disease

in the text there is the description, "The SBBYS variant of Ohdo syndrome (603736) is an allelic disorder with overlapping features."

but it is also not classified:
[Term]
id: OMIM:603736
name: Ohdo Syndrome, Sbbys Variant
is_a: DOID:630 ! genetic disease

so i'd think they would be grouped together in some way.

when i just browse the ontology in protege, if you look under "genetic disease" there are many things that are siblings, but have nearly identical names, suggesting they should be subclasses of a more general disease, but are instead just a big fat flat list. some examples pulled only from A:

Acyl-coa Dehydrogenase * Deficiency Of
Adams-oliver Syndrome
Adrenal Hyperplasia, Congenital
Arthrogryposis * (only 3 of the 6 are classified as subtypes)
Atrial Fibrillation (only 4 of 8 are classified as subtypes)
Atrial Septal Defect (only 4 of 8 are classified as subtypes) ...
in fact, 2257 of the omim diseases in the merged file can't be classified into any upper-level grouping class at all except for a plain "genetic disease", which isn't too helpful.

migrated from monarch-initiative/disease-miner#8

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#6

Möbius, which is also spelled Moebius should not be Mobius.
http://monarchinitiative.org/disease/OMIM_157900
I believe it is wrong in a number of places including NCI, DO, and MESH.

From: monarch-initiative/monarch-app#588

From @nlwashington on May 4, 2015 4:42

Weissenbacher-Zweymüller syndrome has as synonyms Pierre Robin Malformation, Piere-Robin syndrome

PIERRE ROBIN SEQUENCE should be a synonym (should be coming from the OMIM file)

http://tartini.crbs.ucsd.edu/disease/OMIM:261800

also, it is odd to see that OMIM:261800 Pierre Robin Syndrome is a parent of itself (with same OMIM id) (see screenshot).

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#32

From @nlwashington on March 6, 2015 17:41

currently, ids for SNOMED are formed like:

SNOMEDCT_2010_1_31:154999006

is it necessary to code the date/version information into the prefix? shouldn't this just be SNOMEDCT?

also in ticket: https://support.crbs.ucsd.edu/browse/NIF-11604

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#5

From source:

[Term]
id: OMIM:261500
name: Eosinophil Peroxidase Deficiency
def: "Eosinophil peroxidase deficiency is a rare abnormality of eosinophil granulocytes characterized by decreased or absent peroxidase activity and decreased volume of the granule matrix (summary by {9:Romano et al., 1994})." []
synonym: "Eosinophil Peroxidase Deficiency, Partial" EXACT []
synonym: "EOSINOPHIL PEROXIDASE DEFICIENCY; EPXD" EXACT []
synonym: "EPXD" EXACT []
synonym: "Peroxidase and Phospholipid Deficiency 1N Eosinophils" EXACT []
synonym: "Presentey Anomaly" EXACT []
xref: UMLS:C1850000

not showing in MONDO

From @nlwashington on May 20, 2015 3:27

there ought to be a grouping class for all ciliopathies.

http://en.wikipedia.org/wiki/Ciliopathy

i would think that primary ciliary dyskinesia would be one of these, as well as many others. it seems like Meckel Syndrome, Type 3 is one of these, but there is no general group for all ciliopathies (including Bardet-Biedel)

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#35

From @mellybelly on March 13, 2015 17:29

We have
DOID_0080014 Chromosomal disease
and
Orphanet_68335 Chromosomal anomaly

These are currently siblings, but should probably be equivalent classes. Ideally these would be inferred based upon some genetic characterization of the actual chromosomal variation. @nlwashington @mbrush can we make a property that we populate via annotations to autoclassify these in the future? We could curate these when we go through the disease annotations and add hallmark features, etc, in the new annotation tool.

The DO definition is: A genetic disease that has_material_basis_in extra, missing, or re-arranged chromosomes. Orphanet has no definition, but I like the subclasses in that they are based upon specific chromosomes. It should be easy to infer under these types with the above annotation (for example, 'partial deletion of chromosome 1')

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#13

From @mellybelly on May 1, 2015 22:56

These are likely in the UMLS but we'll want them so as to associate drug-disease relationships.
@ShahimEssaid working on NDFRT but maybe can grab the disease graph and ensure we have those classes/xrefs in MonDO

http://evs.nci.nih.gov/ftp1/NDF-RT/

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#31

From @mellybelly on March 25, 2015 16:52

from monarch-app #581

ORPHANET:342 and OMIM_249100 should be merged into the one disease they are. They've got different number of annotations and likely more differences/similarities, but they're the same disease.

In addition there is another OMIM record for FMF which is autosomal dominant.

Be sensible to have these rationally connected.

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#23

From @mellybelly on March 14, 2015 23:25

see required classes in directory
https://github.com/monarch-initiative/human-disease-ontology/tree/master/src/contrib/MesH.contrib

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#15

From @cmungall on March 17, 2015 18:42

How is this coming along?

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#19

From @cmungall on March 19, 2015 5:14

Example: Orphanet:924. We have HPO annotations to this.

This is apparently still in orphanet (on the website), but vanished from ORDO, which means we don't even have a label for it.

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#22

From @cmungall on April 6, 2015 15:58

Something happened in the latest build and these became unmerged.

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#27

From @cmungall on August 8, 2015 19:11

We have:

 / DOID:4 ! disease
  is_a DOID:0080015 ! physical disorder
   is_a DOID:0050567 ! orofacial cleft
    is_a DOID:674 ! cleft palate
     is_a OMIM:119540 ! Cleft Palate *** 

This is confusing for users. We deliberately are non-aggressive in merging OMIMs into DOID; the DO term is typically general, as in this case, with the OMIM sometimes referring to something more specific than it's name suggests, although this is far from clear:

  is_a DOID:674 ! cleft palate 
     is_a OMIM:119540 ! Cleft Palate *** SPECIFIC OR GENERAL? ***
     is_a MESH:C535294 ! Reardon Hall Slaney syndrome
     is_a MESH:C535622 ! Halal syndrome
     is_a MESH:C535774 ! Piepkorn Karp Hickok syndrome
     is_a MESH:C535846 ! Hay Wells syndrome recessive type
     is_a MESH:C535971 ! Coloboma, cleft lip/palate and mental retardation syndrome
     is_a MESH:C536027 ! Martinez Monasterio Pinheiro syndrome
     is_a MESH:C536179 ! Feingold Trainer syndrome
     is_a MESH:C536189 ! Ectrodactyly-cleft lip/palate syndrome
     is_a MESH:C536284 ! Iida Kannari syndrome
     is_a MESH:C536639 ! Schrander-Stumpel Theunissen Hulsmans syndrome
     is_a MESH:C536721 ! Zadik Barak Levin syndrome
     is_a MESH:C537231 ! Samson Viljoen syndrome
     is_a MESH:C537585 ! Seres-Santamaria Arimany Muniz syndrome
     is_a MESH:C537795 ! Baetz-Greenwalt syndrome
     is_a MESH:C537901 ! Bamforth syndrome
     is_a MESH:C537906 ! Baraitser Rodeck Garner syndrome
     is_a MESH:C538126 ! Kuster syndrome
     is_a MESH:C538325 ! Ho Kaufman Mcalister syndrome
     is_a MESH:C563509 ! Schilbach-Rott Syndrome Ocular Hypotelorism, Submucosal Cleft Palate, and Hypospadias
     is_a MESH:C563651 ! Kallmann Syndrome 2 with Cleft Lip or Palate
     is_a MESH:C564442 ! Cleft Palate with Ankyloglossia
     is_a MESH:C565173 ! Coloboma, Uveal, with Cleft Lip and Palate and Mental Retardation
     is_a MESH:C565256 ! Diamond-Blackfan Anemia With Microtia And Cleft Palate
     is_a MESH:C565603 ! Cleft Lip with or without Cleft Palate, Nonsyndromic, 7
     is_a OMIM:303400 ! Cleft palate X-linked
     is_a Orphanet:101023 ! Cleft hard palate
     is_a Orphanet:155878 ! Submucosal cleft palate
     is_a Orphanet:99771 ! Bifid uvula
     is_a Orphanet:99772 ! Cleft velum

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#41

From @mellybelly on June 12, 2015 10:56

For example see:
http://monarchinitiative.org/disease/OMIM_182290
http://monarchinitiative.org/disease/DECIPHER_8

Also see
https://decipher.sanger.ac.uk/syndrome/8#phenotype
for comparison. We have a lot more phenotypes, but we should check we aren't missing annotations

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#37

From @mellybelly on March 14, 2015 21:53

For example:
OMIM_107730 Apolipoprotein B; APOB
subclass of DO:disease and Orphanet_121386 (no label in file, but is a gene from Orphanet apolipoprotein B)

coming from MGI file I think, genes need to be pruned out upstream.

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#14

I think it would be easier for folks to review the outcomes of the report if they can see everything in EXCEL

You could take each string provided, and then in a different column show each ID and each primary label associated with it. then you can also show all the equiv classes and their IDs and labels. In a different column, add the superclass ID and label.

We might try creatively sorting or color coding the same matches for each gene or some such in the excel for easy review.

there are a lot of disease terms from that are children of thing in the bridge.owl file.
1 DECIPHER term
20+ DC terms
approx 100 DC terms
many many MESH terms
Orphanet terms too

From @nlwashington on April 6, 2015 23:4

i was poking around, and saw "Rare systemic or rheumatologic disease" (Orphanet:98023) as high-level disease cluster

i was trying to figure out what these were, since there's no definition for this term.

i came across this page:
http://www.merckmanuals.com/professional/musculoskeletal_and_connective_tissue_disorders/autoimmune_rheumatic_disorders/systemic_sclerosis.html
(in our system, http://tartini.crbs.ucsd.edu/disease/DOID:418)

in which i notice that it is considered a "musculoskeletal and connective tissue disorder". i wonder if this disease or cluster needs to be additionally classified as such? (i think it is connective tissue, but not musculoskeletal)

also i notice that when i query for 'rheumatic disease' (MESH:D012216), no diseases come with it, but it's "parent", which is identical in label except for capitalization, has loads: DOID:1575

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#28

From @cmungall on March 13, 2015 5:11

All OMIMs are genetic. We should ensure each one is classified here, directly or indirectly. Note that it's not clear which orphanets belong under here.

In future we may want to try representing this in a way that doesn't overload isa - e.g. with an etiology field

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#10

From @nlwashington on March 31, 2015 20:43

the readme here still looks identical to the project it was forked from. perhaps it should be revised/rebranded?

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#26

From @nlwashington on March 6, 2015 17:37

please add medgen ids to the uberdisease ontology for xrefs. these will be useful for linking with clinvar data.
http://www.ncbi.nlm.nih.gov/medgen/docs/faq/#rel

also found this possibly relevant:
ftp://ftp.ncbi.nih.gov/pub/medgen/MedGen_HPO_OMIM_Mapping.txt.gz

ftp://ftp.ncbi.nih.gov/gene/DATA/mim2gene_medgen

also see related monarch-initiative/disease-miner#5
and https://support.crbs.ucsd.edu/browse/NIF-11766

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#4

From @cmungall on March 17, 2015 5:18

probably coming from omimclusters, which has its own pipeline for pulling in omim. We should swap in our version

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#18

From @cmungall on July 20, 2015 23:39

It probably makes sense to make this repo standalone rather than a fork, as it currently is

can be done this way
https://help.github.com/articles/duplicating-a-repository/

But ideally we would preserve tickets too

See hints here:
holman/ama#413

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#40

From @mellybelly on March 6, 2015 19:6

It would be great if we could autopopulate any missing intermediate nodes based upon anatomical classification of hallmark phenotypes. Basically, for 'Enlarged hippocampus', you know that hippocampal formation is part of 'pallium' and 'cerebral cortex'. If Enlarged hippocampus is a hallmark for a disease (for some reason in our site I don't see any diseases with mophological abnormality of the hippocampus...hm) then we'd autoclassify that disease under the seeded intermediate nodes 'abnormal pallium phenotype' and 'abnormal cerebral cortex phenotype' or some such.

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#7

From @nicolevasilevsky on May 6, 2015 18:46

I made a test comment and was able to save it, but now I am having issues again. :(

I tried to move one class (achoo syndrome to a child of syndrome) and it gave me an error when I tried to save, then I tried to save again and it told me the ontology had change and would l like to reload, so I said yes.

Then I closed Protege (5, beta 17) and now it won't load the bridge file:

Another question- do I need to open the mondo.obo file before I can open the bridge.owl file? That is what I have been doing.

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#33

From @cmungall on March 13, 2015 6:31

Should we have a DC to cluster the two OMIMs (brown) here?

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#12

From @mellybelly on March 15, 2015 7:11

I'm seeing quite a lot in NCI and in other places that are not in DO or other sources. perhaps we can do some comparisons? eg. align with 'disease or disorder' branch and/or review xrefs in DO?

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#16

From @cmungall on March 17, 2015 4:11

We have 114 Orphanet IDs for which we have HP annotation, but the ID does not exist in the ORDO OWL export.

These appear to correspond to entries with no classification. E.g.

http://www.orpha.net/consor/cgi-bin/OC_Exp.php?lng=en&Expert=139498 'Brachydactyly type D'

Should we try and classify these? E.g. this one would go under DOID:0050581 ! brachydactyly

or

Orphanet:294937 ! Brachydactyly
Orphanet:69028 ! Syndrome with brachydactyly

Copied from original issue: monarch-initiative/deprecated-human-disease-ontology#17