Overview of the Cohort Browser
Last updated
Was this helpful?
Last updated
Was this helpful?
Please note: in order to use Cohort Browser on the Platform, an Apollo License is needed.
Use cohort browser for browsing, visualizing data, and creating cohorts
Access geno-pheno patient records from customer's studies, those of collaborators, and/or public datasets (UK Biobank)
Quickly identify phenotype-specific and/or genotype-specific virtual cohorts for GWAS/PheWAS analyses as well as compare cohorts
Visualize the cohort data to understand the underlying pheno/geno distribution
Limited to 15 tiles overall in dashboard
Limited to 30 columns in Data Preview
Add 1-2 tiles at a time, wait for them to refresh before adding more tiles.
Open Cohort Browser
Select Add tile on the top right corner
Find the characteristic you want as a tile and select "add tile"
Repeat until you select the amount of tiles that you are wanting (up to 15)
Used for more advanced comparisons
Add comparisons by selecting the first filter, then selecting the "+" sign for a secondary field
Then, you will edit in the data field details.
Here is the overview of the 2D plots that you can have:
Steps to Create 2D plots
Open Cohort Browser
Select Add tile on the top right corner
Find the characteristic you are wanting to start with and select it, such as biological sex. This is the same step as adding a regular tile, but you will NOT select Add tile.
Instead, add a secondary field by selecting this next to the second characteristic you are wanting to view, such as Age at Recruitment
You will then have options to change the graph with those parameters. It will look something like this
Then, select the add tile button on the bottom right below the new graph. This will add it to the cohort browser.
In the Cohort section, select the + Add Filter
Search or Select your characteristic
Example: Population Characteristics > Baseline Characteristics > Sex
Click Add Cohort Filter
Make sure "Is Any of" is selected, click on empty field
Select details for the characteristic
Example: selecting female
Your cohort panel will then look like this:
Repeat steps as necessary to filter as needed to create your cohort
By default, filters return all rows belonging to a patient when you query a secondary entity
Option 1: 2 filters at the same "level" but different rows
In this example, we are telling the cohort browser to find patients that have taken the medication ipilimumab OR (and with) the medication stop reason being side effects
This is because of how the filters were added. Essentially, this example is looking at the the medication list and the medication stop reason at the same time.
Option 2: filters on the same row
In this example, we are telling the cohort browser to find patients that have taken the medication ipilimumab AND who stopped taking it due to side effects.
This example, essentially, puts them on the same row and views the parameters together as a pair.
Option 1:
In the example, we use coffee drinkers.
If in the cohort filter we select sex as female AND the coffee consumed is 0, then only females with an instance of 0 for coffee consumed will show.
The differences to note between the and/ or function are highlighted below:
If in the cohort filter we select sex as female OR the coffee consumed is 0, then both females and males with an instance of 0 for coffee consumed will show. This increases the participant size because all the females are included, OR those that have a coffee instance of 0 (could be male OR female).
The differences in this cohort filtering is highlighted below:
Option 2:
In this example, we wanted patients with variants that are common or rare germline OR (and with) variants that are somatic AND have a strong evidence level (A/B).
This example demonstrates combining AND/ OR filters with data elements from a single table so that you can view what you are wanting in a cohort.
In this example, we wanted a patient with a Tumor that has an intermediate or high tumor mutation burden OR (and with) a moderate or high impact mutation in APC, TP53, MSH6, PMS2.
This example is highlighting that you can combine phenotype and genotype filters so that you can customize your cohort further.
You can compare 2 cohorts in order to test hypotheses
To make a control cohort, you can copy the test cohort and change the filters.
For this example, we used coffee drinkers and then everyone else as a comparison
The first cohort is based on coffee type. This cohort was about 17,000 phenotypes.
The second cohort was all 100,000 phenotypes. We add this by selecting on the top by the name of the dataset to add another cohort in.
This allows us to see the 2 of them side by side in the tiles that we have selected. Coffee Drinkers was blue, and all of the 100,00 were purple in the example given.
Once data is ingested, they are available as separate Spark databases. Apollo unifies accessing data in these databases through what's called a dataset.
A dataset can be thought of as a giant multi-omics matrix.
Datasets can be further refined into Cohorts within the Apollo Interface, allowing complex queries across omics type
Need applications for datasets/ cohorts
They exist on the platform in the tool library
This allows for further functionality with Apollo.
These apps allow the use of multiple files at a time (such as a cohort) and process them accordingly.
Some of these can also help process the dataset further as well.
Filtering a Dataset
To create a support ticket if there are technical issues:
Go to the Help header (same section where Projects and Tools are) inside the platform
Select "Contact Support"
Fill in the Subject and Message to submit a support ticket.
See the for how to use these.
