Analysis of differences with different gene to disease sources

Load all associations from multiple sources

The mondo-g2d input spec includes metadata about where to download

gencc
medgen_mim2gene

Note these use heterogeneous IDs, the spec includes normalizers for these

[1]:

from oaklib import get_adapter

[2]:

from oaklib.conf import CONF_DIR_PATH
adapter = get_adapter(CONF_DIR_PATH/ "mondo-g2d-input-spec.yaml")

[3]:

associations = list(adapter.associations())

[4]:

len(associations)

[4]:

[5]:

from linkml_runtime.dumpers import json_dumper

[6]:

import pandas as pd

[7]:

def _as_dict(a):
    obj = json_dumper.to_dict(a)
    obj["subject_prefix"] = a.subject.split(":")[0]
    obj["object_prefix"] = a.object.split(":")[0]
    return obj

[8]:

df = pd.DataFrame([_as_dict(a) for a in associations])
df.to_csv("g2d.tsv", sep="\t", index=False)

[9]:

df

[9]:

	subject	predicate	object	primary_knowledge_source	aggregator_knowledge_source	original_object	original_subject	subject_prefix	object_prefix	subject_label	object_label
0	HGNC:1952	biolink:gene_associated_with_condition	MONDO:0007032	infores:medgen_mim_g2d	infores:medgen_mim_g2d	OMIM:100100	NCBIGene:1131	HGNC	MONDO	NaN	NaN
1	HGNC:29216	biolink:gene_associated_with_condition	MONDO:0024506	infores:medgen_mim_g2d	infores:medgen_mim_g2d	OMIM:100300	NCBIGene:57514	HGNC	MONDO	NaN	NaN
2	HGNC:3690	biolink:gene_associated_with_condition	MONDO:0007037	infores:medgen_mim_g2d	infores:medgen_mim_g2d	OMIM:100800	NCBIGene:2261	HGNC	MONDO	NaN	NaN
3	HGNC:7773	biolink:gene_associated_with_condition	MONDO:0007039	infores:medgen_mim_g2d	infores:medgen_mim_g2d	OMIM:101000	NCBIGene:4771	HGNC	MONDO	NaN	NaN
4	HGNC:3689	biolink:gene_associated_with_condition	MONDO:0007041	infores:medgen_mim_g2d	infores:medgen_mim_g2d	OMIM:101200	NCBIGene:2263	HGNC	MONDO	NaN	NaN
...	...	...	...	...	...	...	...	...	...	...	...
106572	HGNC:12762	NaN	MONDO:0013673	GENCC:000110	infores:gencc	NaN	NaN	HGNC	MONDO	WFS1	Wolfram-like syndrome
106573	HGNC:12762	NaN	MONDO:0013673	GENCC:000102	infores:gencc	NaN	NaN	HGNC	MONDO	WFS1	Wolfram-like syndrome
106574	HGNC:12762	NaN	MONDO:0013673	GENCC:000101	infores:gencc	NaN	NaN	HGNC	MONDO	WFS1	Wolfram-like syndrome
106575	HGNC:29271	NaN	MONDO:0030491	ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/mim2gene_...	infores:hpoa	OMIM:619644	NCBIGene:57169	HGNC	MONDO	ZNFX1	NaN
106576	HGNC:29271	NaN	MONDO:0030491	GENCC:000101	infores:gencc	NaN	NaN	HGNC	MONDO	ZNFX1	immunodeficiency 91 and hyperinflammation

106577 rows × 11 columns

[10]:

df.query("object == 'MONDO:0008426'")

[10]:

	subject	predicate	object	primary_knowledge_source	aggregator_knowledge_source	original_object	original_subject	subject_prefix	object_prefix	subject_label	object_label
699	HGNC:10896	biolink:gene_associated_with_condition	MONDO:0008426	infores:medgen_mim_g2d	infores:medgen_mim_g2d	OMIM:182212	NCBIGene:6497	HGNC	MONDO	NaN	NaN
15644	HGNC:10896	NaN	MONDO:0008426	ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/mim2gene_...	infores:hpoa	OMIM:182212	NCBIGene:6497	HGNC	MONDO	SKI	NaN
15645	HGNC:10896	NaN	MONDO:0008426	http://www.orphadata.org/data/xml/en_product6.xml	infores:hpoa	Orphanet:2462	NCBIGene:6497	HGNC	MONDO	SKI	NaN
15646	HGNC:10896	NaN	MONDO:0008426	GENCC:000101	infores:gencc	NaN	NaN	HGNC	MONDO	SKI	Shprintzen-Goldberg syndrome
15647	HGNC:10896	NaN	MONDO:0008426	GENCC:000112	infores:gencc	NaN	NaN	HGNC	MONDO	SKI	Shprintzen-Goldberg syndrome
...	...	...	...	...	...	...	...	...	...	...	...
104537	HGNC:10896	NaN	MONDO:0008426	GENCC:000111	infores:gencc	NaN	NaN	HGNC	MONDO	SKI	Shprintzen-Goldberg syndrome
104538	HGNC:10896	NaN	MONDO:0008426	GENCC:000104	infores:gencc	NaN	NaN	HGNC	MONDO	SKI	Shprintzen-Goldberg syndrome
104539	HGNC:10896	NaN	MONDO:0008426	GENCC:000106	infores:gencc	NaN	NaN	HGNC	MONDO	SKI	Shprintzen-Goldberg syndrome
104540	HGNC:10896	NaN	MONDO:0008426	GENCC:000110	infores:gencc	NaN	NaN	HGNC	MONDO	SKI	Shprintzen-Goldberg syndrome
104541	HGNC:10896	NaN	MONDO:0008426	GENCC:000102	infores:gencc	NaN	NaN	HGNC	MONDO	SKI	Shprintzen-Goldberg syndrome

98 rows × 11 columns

[11]:

df.query("aggregator_knowledge_source == 'infores:hpoa'")

[11]:

	subject	predicate	object	primary_knowledge_source	aggregator_knowledge_source	original_object	original_subject	subject_prefix	object_prefix	subject_label	object_label
5907	HGNC:16391	NaN	MONDO:0008905	ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/mim2gene_...	infores:hpoa	OMIM:212050	NCBIGene:64170	HGNC	MONDO	CARD9	NaN
5908	HGNC:16391	NaN	MONDO:0008905	http://www.orphadata.org/data/xml/en_product6.xml	infores:hpoa	Orphanet:457088	NCBIGene:64170	HGNC	MONDO	CARD9	NaN
5911	HGNC:21066	NaN	MONDO:0009544	ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/mim2gene_...	infores:hpoa	OMIM:248000	NCBIGene:51256	HGNC	MONDO	TBC1D7	NaN
5915	HGNC:14313	NaN	MONDO:0033485	ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/mim2gene_...	infores:hpoa	OMIM:617895	NCBIGene:28981	HGNC	MONDO	IFT81	NaN
5916	HGNC:6742	NaN	MONDO:0014693	ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/mim2gene_...	infores:hpoa	OMIM:616564	NCBIGene:8216	HGNC	MONDO	LZTR1	NaN
...	...	...	...	...	...	...	...	...	...	...	...
106562	HGNC:12731	NaN	MONDO:0010294	ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/mim2gene_...	infores:hpoa	OMIM:300299	NCBIGene:7454	HGNC	MONDO	WAS	NaN
106563	HGNC:12731	NaN	MONDO:0010294	http://www.orphadata.org/data/xml/en_product6.xml	infores:hpoa	Orphanet:86788	NCBIGene:7454	HGNC	MONDO	WAS	NaN
106567	HGNC:12762	NaN	MONDO:0013673	ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/mim2gene_...	infores:hpoa	OMIM:614296	NCBIGene:7466	HGNC	MONDO	WFS1	NaN
106568	HGNC:12762	NaN	MONDO:0013673	http://www.orphadata.org/data/xml/en_product6.xml	infores:hpoa	Orphanet:411590	NCBIGene:7466	HGNC	MONDO	WFS1	NaN
106575	HGNC:29271	NaN	MONDO:0030491	ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/mim2gene_...	infores:hpoa	OMIM:619644	NCBIGene:57169	HGNC	MONDO	ZNFX1	NaN

40613 rows × 11 columns

[12]:

df.groupby(["subject", "object"])["aggregator_knowledge_source"].describe()

[12]:

		count	unique	top	freq
subject	object
HGNC:10001	MONDO:0007781	4	2	infores:hpoa	2
HGNC:10004	MONDO:0012033	50	3	infores:gencc	35
HGNC:10006	MONDO:0008493	17	3	infores:hpoa	8
HGNC:10006	MONDO:0019107	10	3	infores:hpoa	6
HGNC:10008	MONDO:0019107	4	2	infores:hpoa	2
...	...	...	...	...	...
NCBIGene:105259599	MONDO:0020796	2	2	infores:medgen_mim_g2d	1
NCBIGene:109580095	MONDO:0013517	2	2	infores:medgen_mim_g2d	1
NCBIGene:111365204	MONDO:0007630	2	2	infores:medgen_mim_g2d	1
NCBIGene:111365204	MONDO:0010932	2	2	infores:medgen_mim_g2d	1
NCBIGene:7467	MONDO:0008684	2	2	infores:medgen_mim_g2d	1

15696 rows × 4 columns

[13]:

df.groupby(["subject_prefix", "object_prefix"])["aggregator_knowledge_source"].describe()

[13]:

		count	unique	top	freq
subject_prefix	object_prefix
HGNC	MONDO	105710	3	infores:gencc	60057
	OMIM	719	2	infores:hpoa	405
	Orphanet	132	1	infores:hpoa	132
NCBIGene	MONDO	16	2	infores:medgen_mim_g2d	8

[14]:

df.groupby(["subject_prefix", "object_prefix", "aggregator_knowledge_source"]).describe()

[14]:

			subject				predicate				object		...	original_subject		subject_label				object_label
			count	unique	top	freq	count	unique	top	freq	count	unique	...	top	freq	count	unique	top	freq	count	unique	top	freq
subject_prefix	object_prefix	aggregator_knowledge_source
HGNC	MONDO	infores:gencc	60057	4753	HGNC:3690	246	0	0	NaN	NaN	60057	6155	...	NaN	NaN	60057	4753	FGFR3	246	60057	6151	retinitis pigmentosa	229
		infores:hpoa	40068	5128	HGNC:2200	130	0	0	NaN	NaN	40068	7527	...	NCBIGene:1280	130	40068	5128	COL2A1	130	0	0	NaN	NaN
		infores:medgen_mim_g2d	5585	3933	HGNC:2200	14	5585	1	biolink:gene_associated_with_condition	5585	5585	5093	...	NCBIGene:1280	14	0	0	NaN	NaN	0	0	NaN	NaN
	OMIM	infores:hpoa	405	385	HGNC:4398	2	0	0	NaN	NaN	405	404	...	NCBIGene:2783	2	405	385	GNB2	2	0	0	NaN	NaN
	OMIM	infores:medgen_mim_g2d	314	305	HGNC:18420	2	314	1	biolink:gene_associated_with_condition	314	314	314	...	NCBIGene:29072	2	0	0	NaN	NaN	0	0	NaN	NaN
	Orphanet	infores:hpoa	132	111	HGNC:7577	2	0	0	NaN	NaN	132	83	...	NCBIGene:4625	2	132	111	MYH7	2	0	0	NaN	NaN
NCBIGene	MONDO	infores:hpoa	8	5	NCBIGene:105259599	3	0	0	NaN	NaN	8	8	...	NaN	NaN	8	1	-	8	0	0	NaN	NaN
NCBIGene	MONDO	infores:medgen_mim_g2d	8	5	NCBIGene:105259599	3	8	1	biolink:gene_associated_with_condition	8	8	8	...	NaN	NaN	0	0	NaN	NaN	0	0	NaN	NaN

8 rows × 32 columns

[15]:

df.sort_values(["subject", "object"])

[15]:

	subject	predicate	object	primary_knowledge_source	aggregator_knowledge_source	original_object	original_subject	subject_prefix	object_prefix	subject_label	object_label
16423	HGNC:10001	NaN	MONDO:0007781	ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/mim2gene_...	infores:hpoa	OMIM:145500	NCBIGene:8490	HGNC	MONDO	RGS5	NaN
16424	HGNC:10001	NaN	MONDO:0007781	GENCC:000106	infores:gencc	NaN	NaN	HGNC	MONDO	RGS5	essential hypertension, genetic
47487	HGNC:10001	NaN	MONDO:0007781	ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/mim2gene_...	infores:hpoa	OMIM:145500	NCBIGene:8490	HGNC	MONDO	RGS5	NaN
47488	HGNC:10001	NaN	MONDO:0007781	GENCC:000106	infores:gencc	NaN	NaN	HGNC	MONDO	RGS5	essential hypertension, genetic
2684	HGNC:10004	biolink:gene_associated_with_condition	MONDO:0012033	infores:medgen_mim_g2d	infores:medgen_mim_g2d	OMIM:608415	NCBIGene:8787	HGNC	MONDO	NaN	NaN
...	...	...	...	...	...	...	...	...	...	...	...
7192	NCBIGene:111365204	NaN	MONDO:0007630	ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/mim2gene_...	infores:hpoa	OMIM:136550	NaN	NCBIGene	MONDO	-	NaN
2108	NCBIGene:111365204	biolink:gene_associated_with_condition	MONDO:0010932	infores:medgen_mim_g2d	infores:medgen_mim_g2d	OMIM:600790	NaN	NCBIGene	MONDO	NaN	NaN
9063	NCBIGene:111365204	NaN	MONDO:0010932	ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/mim2gene_...	infores:hpoa	OMIM:600790	NaN	NCBIGene	MONDO	-	NaN
817	NCBIGene:7467	biolink:gene_associated_with_condition	MONDO:0008684	infores:medgen_mim_g2d	infores:medgen_mim_g2d	OMIM:194190	NaN	NCBIGene	MONDO	NaN	NaN
11624	NCBIGene:7467	NaN	MONDO:0008684	ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/mim2gene_...	infores:hpoa	OMIM:194190	NaN	NCBIGene	MONDO	-	NaN

106577 rows × 11 columns

[16]:

df.groupby("primary_knowledge_source")[["subject", "object"]].count()

[16]:

	subject	object
primary_knowledge_source
GENCC:000101	9974	9974
GENCC:000102	5687	5687
GENCC:000104	4952	4952
GENCC:000105	1365	1365
GENCC:000106	9101	9101
GENCC:000107	1145	1145
GENCC:000108	767	767
GENCC:000110	16407	16407
GENCC:000111	3624	3624
GENCC:000112	6918	6918
GENCC:000113	72	72
GENCC:000114	45	45
ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/mim2gene_medgen	20368	20368
http://www.orphadata.org/data/xml/en_product6.xml	20245	20245
infores:medgen_mim_g2d	5907	5907

[25]:

import matplotlib.pyplot as plt
from matplotlib_venn import venn2, venn3

s1 = "infores:gencc"
s2 = "infores:medgen_mim_g2d"
s3 = "infores:hpoa"
src_col = "aggregator_knowledge_source"

def make_venn(cols, venn_df=None):
    if venn_df is None:
        venn_df = df
    source1_tuples = set(venn_df[venn_df[src_col] == s1][cols].apply(tuple, axis=1))
    source2_tuples = set(venn_df[venn_df[src_col] == s2][cols].apply(tuple, axis=1))
    source3_tuples = set(venn_df[venn_df[src_col] == s3][cols].apply(tuple, axis=1))
    venn3([source1_tuples, source2_tuples, source3_tuples], set_labels=(s1, s2, s3))
    plt.title(f"Venn Diagram for {cols}")
    plt.show()

# Identifying unique (a, b) tuples for each source
#source1_tuples = set(df[df[src_col] == s1][['subject', 'object']].apply(tuple, axis=1))
#source2_tuples = set(df[df[src_col] == s2][['subject', 'object']].apply(tuple, axis=1))

# Creating the Venn diagram
#venn2([source1_tuples, source2_tuples], set_labels=(s1, s2))

#plt.title("Venn Diagram for gene-disease tuples")
#plt.show()

make_venn(["subject", "object"])

../../../_images/examples_Ontologies_MONDO_Disease2GeneAnalysis_18_0.png

[26]:

make_venn(["subject"])

../../../_images/examples_Ontologies_MONDO_Disease2GeneAnalysis_19_0.png

[27]:

make_venn(["object"])

../../../_images/examples_Ontologies_MONDO_Disease2GeneAnalysis_20_0.png

[28]:

diseases = set(df["object"])
len(diseases)

[28]:

[29]:

from oaklib.datamodels.vocabulary import IS_A

redundant = set()
roll_up_map = {}
for d in diseases:
    ancs = adapter.ancestors(d, predicates=[IS_A], reflexive=False)
    ixn = diseases.intersection(ancs)
    redundant.update(ixn)
    if ixn:
        roll_up_map[d] = list(ixn)
    else:
        roll_up_map[d] = [d]

len(redundant)

[29]:

[30]:

df['object_rollup'] = df['object'].map(lambda x: tuple(sorted(roll_up_map[x])))
df["object_rollup"]

[30]:

0                       (MONDO:0005240,)
1         (MONDO:0005172, MONDO:0007034)
2         (MONDO:0005172, MONDO:0005516)
3         (MONDO:0005071, MONDO:0021061)
4         (MONDO:0005172, MONDO:0015469)
                       ...
106572                  (MONDO:0013673,)
106573                  (MONDO:0013673,)
106574                  (MONDO:0013673,)
106575    (MONDO:0005046, MONDO:0021094)
106576    (MONDO:0005046, MONDO:0021094)
Name: object_rollup, Length: 106577, dtype: object

[31]:

make_venn(["subject", "object_rollup"])

../../../_images/examples_Ontologies_MONDO_Disease2GeneAnalysis_24_0.png

[15]:

cross_tab = pd.crosstab(df[df['primary_knowledge_source']=="infores:gencc"]['object'], df[df['primary_knowledge_source']=="infores:medgen_mim_g2d"]['object'])

[16]:

cross_tab

[16]:

object
object

[11]:

gencc = df[df['primary_knowledge_source']=="infores:gencc"]
gencc

[11]:

	subject	predicate	object	primary_knowledge_source
5837	HGNC:10896	NaN	MONDO:0008426	infores:gencc
5838	HGNC:16636	NaN	MONDO:0008233	infores:gencc
5839	HGNC:16636	NaN	MONDO:0007308	infores:gencc
5840	HGNC:17939	NaN	MONDO:0044322	infores:gencc
5841	HGNC:11071	NaN	MONDO:0014572	infores:gencc
...	...	...	...	...
16089	HGNC:17625	NaN	MONDO:0100284	infores:gencc
16090	HGNC:13128	NaN	MONDO:0100148	infores:gencc
16091	HGNC:13156	NaN	MONDO:0100284	infores:gencc
16092	HGNC:29046	NaN	MONDO:0030695	infores:gencc
16093	HGNC:6827	NaN	MONDO:0030770	infores:gencc

10257 rows × 4 columns

[12]:

medgen = df[df['primary_knowledge_source']=="infores:medgen_mim_g2d"]
medgen

[12]:

	subject	predicate	object	primary_knowledge_source
0	HGNC:1952	biolink:gene_associated_with_condition	MONDO:0007032	infores:medgen_mim_g2d
1	HGNC:29216	biolink:gene_associated_with_condition	MONDO:0024506	infores:medgen_mim_g2d
2	HGNC:3690	biolink:gene_associated_with_condition	MONDO:0007037	infores:medgen_mim_g2d
3	HGNC:7773	biolink:gene_associated_with_condition	MONDO:0007039	infores:medgen_mim_g2d
4	HGNC:3689	biolink:gene_associated_with_condition	MONDO:0007041	infores:medgen_mim_g2d
...	...	...	...	...
5832	HGNC:11016	biolink:gene_associated_with_condition	OMIM:620306	infores:medgen_mim_g2d
5833	HGNC:10485	biolink:gene_associated_with_condition	OMIM:620310	infores:medgen_mim_g2d
5834	HGNC:15979	biolink:gene_associated_with_condition	OMIM:620311	infores:medgen_mim_g2d
5835	HGNC:19946	biolink:gene_associated_with_condition	OMIM:620316	infores:medgen_mim_g2d
5836	HGNC:21839	biolink:gene_associated_with_condition	OMIM:620319	infores:medgen_mim_g2d

5837 rows × 4 columns

[13]:

pd.crosstab(df)

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
Cell In[13], line 1
----> 1 pd.crosstab(df)

TypeError: crosstab() missing 1 required positional argument: 'columns'

[ ]:

!pip install matplotlib-venn

[ ]: