Navigating data ontologies

Note

This notebook uses the legacy AlphaGenome 1 API (alphagenome.models.v1). For AlphaGenome 2 (alphagenome.models.v2), see 1. Choosing a condition and Migrating from AlphaGenome 1.

Tip

Open this tutorial in Google Colab for interactive viewing.

# @title Install AlphaGenome

# @markdown Run this cell to install AlphaGenome.
from IPython.display import clear_output
! pip install alphagenome
clear_output()

Setup and imports

from alphagenome import colab_utils
from alphagenome.models.v1 import dna_client
from google.colab import data_table
import pandas as pd

data_table.enable_dataframe_formatter()

Interactively view output metadata

First, we load the model.

dna_model = dna_client.create(colab_utils.get_api_key())

output_metadata = dna_model.output_metadata(
    dna_client.Organism.HOMO_SAPIENS
).concatenate()

Click Filter on the upper right hand side of the interactive dataframe and type a cell or tissue name like “brain” into the Search by all fields box to find the ontology_curie term corresponding to a tissue and output type of interest:

output_metadata

	name	strand	Assay title	ontology_curie	biosample_name	biosample_type	biosample_life_stage	data_source	endedness	genetically_modified	nonzero_mean	output_type	gtex_tissue	histone_mark	transcription_factor
0	CL:0000084 ATAC-seq	.	ATAC-seq	CL:0000084	T-cell	primary_cell	adult	encode	paired	False	0.739741	OutputType.ATAC	NaN	NaN	NaN
1	CL:0000100 ATAC-seq	.	ATAC-seq	CL:0000100	motor neuron	in_vitro_differentiated_cells	adult	encode	paired	False	0.273136	OutputType.ATAC	NaN	NaN	NaN
2	CL:0000236 ATAC-seq	.	ATAC-seq	CL:0000236	B cell	primary_cell	adult	encode	paired	False	4.700081	OutputType.ATAC	NaN	NaN	NaN
3	CL:0000623 ATAC-seq	.	ATAC-seq	CL:0000623	natural killer cell	primary_cell	adult	encode	paired	False	0.938715	OutputType.ATAC	NaN	NaN	NaN
4	CL:0000624 ATAC-seq	.	ATAC-seq	CL:0000624	CD4-positive, alpha-beta T cell	primary_cell	adult	encode	paired	False	4.365206	OutputType.ATAC	NaN	NaN	NaN
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
7	ENCSR182QNJ	-	PRO-cap	EFO:0001099	Caco-2	cell_line	NaN	encode	NaN	False	14.002803	OutputType.PROCAP	NaN	NaN	NaN
8	ENCSR740IPL	-	PRO-cap	EFO:0002067	K562	cell_line	NaN	encode	NaN	False	15.765458	OutputType.PROCAP	NaN	NaN	NaN
9	ENCSR797DEF	-	PRO-cap	EFO:0002819	Calu3	cell_line	NaN	encode	NaN	False	12.281321	OutputType.PROCAP	NaN	NaN	NaN
10	ENCSR801ECP	-	PRO-cap	CL:0002618	endothelial cell of umbilical vein	primary_cell	NaN	encode	NaN	False	13.973692	OutputType.PROCAP	NaN	NaN	NaN
11	ENCSR860TYZ	-	PRO-cap	EFO:0001200	MCF 10A	cell_line	NaN	encode	NaN	False	8.229502	OutputType.PROCAP	NaN	NaN	NaN

5563 rows × 15 columns

How many tracks are there per output type?

# Count human tracks
human_tracks = (
    dna_model.output_metadata(dna_client.Organism.HOMO_SAPIENS)
    .concatenate()
    .groupby('output_type')
    .size()
    .rename('# Human tracks')
)

# Count mouse tracks
mouse_tracks = (
    dna_model.output_metadata(dna_client.Organism.MUS_MUSCULUS)
    .concatenate()
    .groupby('output_type')
    .size()
    .rename('# Mouse tracks')
)

pd.concat([human_tracks, mouse_tracks], axis=1).astype(pd.Int64Dtype())

	# Human tracks	# Mouse tracks
output_type
OutputType.ATAC	167	18
OutputType.CAGE	546	188
OutputType.DNASE	305	67
OutputType.RNA_SEQ	667	173
OutputType.CHIP_HISTONE	1116	183
OutputType.CHIP_TF	1617	127
OutputType.SPLICE_SITES	4	4
OutputType.SPLICE_SITE_USAGE	734	180
OutputType.SPLICE_JUNCTIONS	367	90
OutputType.CONTACT_MAPS	28	8
OutputType.PROCAP	12	<NA>

Note that PROCAP outputs are not available for mouse.