Query Guide

KG query guide

Other relevant docs

Access

Refer to the Access guide for instructions on accessing the Cell Type Knowledge Graph.

Useful MATCH clauses

Clause to find author annotated cell sets for a specific dataset:

MATCH (cc:Cell_cluster)-[:has_source]->(ds) 
WHERE ds.publication = ['https://doi.org/10.1038/s41586-021-03852-1']

Alternatively, one could use ds.title or ds.collection as a restriction, these take dataset titles and colleciton URLs for CxG respectively.

Clause to match cell sets to CL annotations:

MATCH p=(c:Class:Cell)<-[:composed_primarily_of]-(s1:Cell_cluster)

Using CL structure to find annotations with some specified cell type or its subclasses

MATCH p=(superclass:Cell)<-[:SUBCLASSOF*0..]-(c:Class:Cell)<-[:composed_primarily_of]-(s1:Cell_cluster)

Clause to finds subsets (subclusters) of cell sets

MATCH (s1)<-[:subcluster_of*..]-(s2)

Clause to find the proportion of cells in a cell set from specific tissue

MATCH (cc:Cell_cluster {label: 'Enterocyte' )-[t:tissue]->(anat)
RETURN t.percentage, anat.label, anat.short_form

Putting it all together

Query to find the proportion of cells by tissue on a specific annotation

This was motivated by finding an annotation:

Author_cell_type: 'Enterocyte'; cell_type: 'enterocyte of colon'

Question: Does the tissue origin justify annotating this as a cell type of the colon.

MATCH (cc:Cell_cluster)-[:has_source]->(ds { publication: ['https://doi.org/10.1038/s41586-021-03852-1']}) 
MATCH p=(c:Cell { label: 'enterocyte of colon'})<-[:composed_primarily_of]
-(cc:Cell_cluster{label: 'Enterocyte'})-[t:tissue]->(anat)
RETURN anat.label, t.percentage[0]

anat.label	t.percentage[0]
ileum	58.96
duodenum	25.89
large intestine	1.21
jejunum	6.06
mesenteric lymph node	0.01
small intestine	6.96
colon	0.9

Conclusion: >97% if cells are from the small intestine so this annotation is incorrect.

For a specific dataset, find author annotations that are more granular than the CxG CL annotation

MATCH (s1:Cell_cluster)-[:has_source]->(ds) 
WHERE ds.publication = ['https://doi.org/10.1038/s41586-021-03852-1']
MATCH p=(c:Class:Cell)<-[:composed_primarily_of]-(s1)<-[:subcluster_of*..]-(s2)
RETURN p

For all datasets, find author annotations taht are more granular than the CxG CL annotation, and where the CL term is a leaf node

MATCH p=(c:Class:Cell)<-[:composed_primarily_of]-(s1)<-[:subcluster_of*..]-(s2) where not (c)<-[:SUBCLASSOF]-() return p

Find leaf node CL terms with nested cell sets underneath

Example results from Sikemma:

** Query to find General term used for specific class**

e.g. T-Cell here:

MATCH p=(c:Class:Cell)<-[:composed_primarily_of]-(s1:Cluster) where (c)<-[:SUBCLASSOF]-() and not (s1)<-[:subcluster_of]-() return p

c.label	s1.label	More specific term needed*	CL term to use	Notes
serous secreting cell	SMG serous (nasal)	Y		SMG = submucosal gland. We have no nasal SMG serous term.
tracheobronchial goblet cell	Goblet (subsegmental)	?		Need to check if could be more precise
tracheobronchial serous cell	SMG serous (bronchial)	n	serous secreting cell of bronchus submucosal gland
bronchial goblet cell	Goblet (bronchial)	n
CD4-positive, alpha-beta T cell	CD4 T cells	n
mucus secreting cell	SMG mucous	n	mucus secreting cell of bronchus submucosal gland
ciliated columnar cell of tracheobronchial tree	Multiciliated (non-nasal)	n
dendritic cell	Migratory DCs	y
lung macrophage	Interstitial Mph perivascular	y		"perivascular macrophage" is currently brain specific! We also have lung interstitial macrophage
plasmacytoid dendritic cell	Plasmacytoid DCs	n
smooth muscle cell	SM activated stress response	y?
epithelial cell of alveolus of lung	AT0	Y
fibroblast	Subpleural fibroblasts	Y
tracheobronchial smooth muscle cell	Smooth muscle	N
epithelial cell of lower respiratory tract	pre-TB secretory	Y		TB = tracheobronchial
CD8-positive, alpha-beta T cell	CD8 T cells	N
T cell	T cells proliferating	N
conventional dendritic cell	DC1	N
brush cell of trachebronchial tree	Tuft	N

*quick assessment - may not be 100% acurate