Synonym and disease gaps in EFO

[d0choa]

As part of an Open Targets project in collaboration with EPMC, we have processed the entire corpus to detect diseases or phenotypes using NER. In a subsequent step, we ground each label to its corresponding term in EFO using all available names and synonyms.

This process has identified several gaps in EFO. Labels that are frequently used in the literature, but are not present in EFO either in the form of terms or as synonyms.

I have curated the first set of highly frequent labels for your consideration and divided them into the next 2 groups:

1. Labels with available EFO terms for their inclusion as synonyms - 98 terms accounting for 4.7M literature references in up to 897k publications.

Label occurrences	Distinct PMIDs	Label	EFO name	EFO ID
281184	29040	T2DM	type II diabetes mellitus	EFO_0001360
214329	23081	BC	breast cancer	MONDO_0007254
213784	19137	NAFLD	non-alcoholic fatty liver disease	EFO_0003095
168751	15452	T2D	type II diabetes mellitus	EFO_0001360
166389	21443	UC	ulcerative colitis	EFO_0000729
153989	10065	MetS	metabolic syndrome	EFO_0000195
140109	21681	AMI	acute myocardial infarction	EFO_0008583
137843	16523	MDD	major depressive disorder	MONDO_0002009
126968	43577	HIV	HIV infection	EFO_0000764
122979	9840	TNBC	Triple-negative breast cancer	EFO_0005537
115053	12851	VTE	venous thromboembolism	EFO_0004286
107246	8324	PDAC	pancreatic ductal adenocarcinoma	EFO_0002517
104390	23693	IR	insulin resistance	EFO_0002614
90575	9256	T1D	type I diabetes mellitus	EFO_0001359
88229	8424	OSCC	oral squamous cell carcinoma	EFO_0000199
85396	20998	SCC	squamous cell carcinoma	EFO_0000707
70477	38258	autoimmunity	autoimmune disease	EFO_0005140
65469	8023	OS	osteosarcoma	EFO_0000637
64799	11736	rabies	Rhabdoviridae infectious disease	EFO_0007469
60836	10617	CIN	cervical intraepithelial neoplasia	MONDO_0022394
58922	16418	HBV infection	hepatitis B virus infection	EFO_0004197
57311	10697	AR	allergic rhinitis	EFO_0005854
52704	6803	T1DM	type I diabetes mellitus	EFO_0001359
51090	4985	CDI	clostridium difficile infection	EFO_0009130
50716	8927	OC	ovarian cancer	MONDO_0008170
48762	12803	H. pylori infection	Helicobacter pylori infectious disease	EFO_1000961
44065	6972	PTC	papillary thyroid carcinoma	EFO_0000641
43628	4463	EOC	ovarian carcinoma	EFO_0001075
42817	8579	CS	Cowden syndrome	Orphanet_201
42669	4719	VL	visceral Leishmaniasis	EFO_0005045
42328	10782	cholera	Vibrio infectious disease	EFO_1001235
42163	6352	CHB	Congenital heart block	Orphanet_60041
42163	6352	CHB	chronic hepatitis B virus infection	EFO_0004239
41825	10433	SAH	subarachnoid hemorrhage	EFO_0000713
41807	12310	HPV infection	human papilloma virus infection	EFO_0001668
41596	28921	neurodegenerative disorders	neurodegenerative disease	EFO_0005772
39432	11808	CoV-2 infection	COVID-19	MONDO_0100096
38651	14778	iron deficiency	Iron deficiency anemia	HP_0001891
37339	22242	hepatitis C	hepatitis C virus infection	EFO_0003047
36992	6792	HS	hidradenitis suppurativa	EFO_1000710
36992	6792	HS	hippocampal sclerosis of aging	EFO_0005678
36992	6792	HS	histiocytic sarcoma	MONDO_0019479
36913	4844	PTB	pulmonary tuberculosis	EFO_1000049
35904	20295	cognitive dysfunction	cognitive disorder	EFO_1001457
35279	4599	GD	Gaucher disease	Orphanet_355
35117	9776	DVT	deep vein thrombosis	EFO_0003907
34963	5317	LBP	Low back pain	HP_0003419
34809	6760	NET	neuroendocrine neoplasm	EFO_1001901
33812	4431	AOM	Otitis media	EFO_0004992
32514	2831	IgAN	IGA glomerulonephritis	EFO_0004194
30655	9762	SCI	Spinal cord injury	EFO_1001919
29510	17290	renal dysfunction	impaired renal function disease	MONDO_0001343
29066	11106	HCV infection	hepatitis C virus infection	EFO_0003047
28706	13395	hyperglycaemia	Hyperglycemia	HP_0003074
25718	5123	HH	hypogonadotropic hypogonadism	MONDO_0018555
25538	4379	IE	infective endocarditis	MONDO_0000565
22675	3172	OAB	overactive bladder	EFO_1000781
22648	3292	EAC	Esophageal adenocarcinoma	EFO_0000478
22168	5702	EA	Esophageal atresia	HP_0002032
22096	4304	VAP	Ventilator-associated pneumonia	EFO_1001865
22092	13932	angina	angina	EFO_0003913
21712	2897	FM	fibromyalgia	EFO_0005687
21028	2332	HAT	human african trypanosomiasis	EFO_0005225
19964	4562	TLE	temporal lobe epilepsy	EFO_0000773
19886	2609	MCC	Merkel cell skin cancer	EFO_1001471
19709	3635	SUD	substance-related disorder	MONDO_0002494
19667	2671	UI	Urinary incontinence	HP_0000020
19502	1920	SCZ	schizophrenia	EFO_0000692
19167	6692	malaria infection	malaria	EFO_0001068
18310	1482	DKD	diabetic nephropathy	EFO_0000401
18284	3163	AUD	alcohol abuse	MONDO_0002046
17895	1616	GBC	gallbladder cancer	MONDO_0005411
17857	2371	ALF	Acute hepatic failure	HP_0006554
17335	6266	cardiac fibrosis	Myocardial fibrosis	HP_0001685
17257	2258	MPM	malignant pleural mesothelioma	EFO_0000770
17176	12216	psychiatric illness	psychiatric disorder	MONDO_0002025
17157	10823	cardiac dysfunction	heart failure	EFO_0003144
17053	14219	neurodegenerative disorder	neurodegenerative disease	EFO_0005772
16992	4000	RSV infection	Respiratory Syncytial Virus Infection	EFO_1001413
16484	3218	BA	asthma	EFO_0000270
16484	3218	BA	Biliary atresia	Orphanet_30391
16409	1544	LSCC	laryngeal squamous cell carcinoma	EFO_0006352
15810	2621	GHD	Growth hormone deficiency	MONDO_0000050
15410	2728	PKD	polycystic kidney disease	MONDO_0020642
15283	3964	PsA	psoriatic arthritis	EFO_0003778
15091	9869	coronavirus disease	coronavirus infectious disease	EFO_0007224
15026	1927	LUAD	lung adenocarcinoma	EFO_0000571
14881	3878	ATL	adult T-cell leukemia/lymphoma	MONDO_0019471
14773	1609	OLP	oral lichen planus	EFO_0008517
14344	1197	HZ	Herpes Zoster	EFO_0006510
12809	2395	DENV infection	dengue disease	EFO_0005547
12746	4832	haemophilia	hemophilia	Orphanet_448
12681	5664	influenza virus infection	influenza	EFO_0007328
12661	5009	TB infection	tuberculosis	Orphanet_3389
12572	6780	reflux	gastroesophageal reflux disease	EFO_0003948
11836	2856	T. gondii infection	toxoplasmosis	EFO_0007517
10078	1630	FGR	fetal growth restriction	EFO_0000495
9843	4926	POP	pelvic organ prolapse	EFO_0004710

2. Labels for which I didn't find an equivalent EFO term and might require imports or further curation - 35 terms accounting for >1M occurrences in up to 358k publications.

Label occurrences	Distinct PMIDs	Label	Likely label	Likely term	Note
97169	14552	MCI	Mild cognitive impairment		To import??? Related to HP_0100543
79385	28194	psychological distress			?? to import. Related to EFO_0009095
59747	19788	viremia			?? to import -> HP_0020071
49764	28824	skin lesions			??Import e.g. NCIT_C158524
40587	6834	CRS	Cytokine Release Syndrome		?? Import -> NCIT_C78251
40587	6834	CRS	cytoreductive surgery		?? Import -> NCIT_C132068
39633	4823	POAG	Primary open angle glaucoma		?? Import -> Related to EFO:0004190 and EFO:1001506
37077	18077	renal impairment			?? import -> NCIT_C114592
36992	6792	HS	Hemorrhagic Stroke		?? import
33985	4651	CRPC	Castration-Resistant Prostate Carcinoma		?? import -> Castration-Resistant Prostate Carcinoma - NCIT:C130234
32325	4425	CIA	Chemotherapy-induced anemia		?? import
32274	23047	arterial hypertension			?? import
30684	9132	TBI	traumatic brain injury		?? import --> Related to EFO_0011023
28954	3111	HFpEF	heart failure with preserved ejection fraction		?? import
28137	3078	ILI	influenza-like illness		?? import related to AEOAE:0000100 and SNOMED:95891005
26715	17171	impaired glucose tolerance	impaired glucose tolerance	HP_0040270	?? import
24548	8775	anaphylaxis	anaphylaxis	NCIT_C107101	?? Import
23507	15313	septicemia	septicemia	NCIT_C3364	?? import
22234	4692	varicocele	varicocele	MONDO_0001498	?? import
21527	13705	nosocomial infections	nosocomial infection	MONDO_0043544	?? Import
20607	10948	effusion	effusion	NCIT_C3003	?? to import
19929	3281	alexithymia	alexithymia	MONDO_0000661	?? import
19723	9249	miscarriage	Spontaneous abortion	HP_0005268	?? import
18940	9596	hepatic encephalopathy	Hepatic Encephalopathy	NCIT_C79596	?? import
18654	5467	IGT	impaired glucose tolerance	HP_0040270	?? import - Mentioned before as separate mention
17820	12916	left ventricular dysfunction	left ventricular dysfunction	NCIT_C50629	?? to import
17277	8581	dyslipidaemia	Dyslipidemia	HP_0003611	?? to import
17094	2837	CAC	coronary artery calcium measurement		?? import SNOMED450734004
17047	10685	wound infection	wound infection	NCIT_C45234	?? import
16412	7907	endotoxemia	Endotoxemia	OMIT_0019490	?? import
15793	7448	hemorrhagic stroke	Hemorrhagic Stroke		?? import - see more above (HS)
15737	12087	traumatic brain injury	traumatic brain injury		?? import see more above TBI --> Related to EFO_0011023
15214	7488	delusions	Delusions	HP_0000746	?? import
15194	4733	SSI	Surgical Site Infection	NCIT_C112019	?? import
10289	3744	gonorrhoea			?? import

Happy to provide more detail about our findings.

[paolaroncaglia]

Great extraction + curation effort! Some thoughts to optimise incorporation of results, all focused on the first list (suggested synonyms):

• [ ] Acronyms can be very useful when they're quite unique (e.g. T2DM for type II diabetes mellitus), but highly ambiguous when they are very short (e.g. BC for breast cancer; HS is suggested as synonym of 3 different entities, plus 1 in the second list; and more), with nuances in between. I'd recommend not to add 2- or 3-letter acronyms as exact synonyms, but rather broad or related (note for EFO editors: check if the "abbreviation" annotation property is used in EFO; consider importing if not).
• [ ] Synonyms suggested for addition to Mondo or HP terms should be requested to the source ontologies.
• [ ] 'angina' looks like a duplicate.

[d0choa]

Regarding the first point, it's quite common for dictionaries to contain ambiguous synonyms. For example, we deal with gene/protein dictionaries in which some synonyms (e.g. p55) can refer to multiple instances. That's not a problem of the ontology/dictionary, which captures a reality on the semantic space.

Not including ambiguous synonyms will prevent users to ground labels to EFO. Our take is that the problem of de-ambiguating synonyms or acronyms should be a downstream process of specific applications not a feature of the ontology. This is an already widespread issue in EFO and other dictionaries so we are setting up systems to handle them. More info.

On a separate note, separating acronyms from synonyms would be highly beneficial for us as well. I know this has been mentioned in the past, so I'm just reiterating that we will have an application for it.

[zoependlington]

I have added all EFO/Orphanet acronyms as synonyms for now until we find a better way to represent these separately.

I have also imported all HP and Mondo terms in the second table into EFO. However, currently we are not able to import NCIT and OMIT, so if these and any without a suggested term are required we will need a new term request for these. I can set up a template for a bulk request of these if needed.