← All guides

Trajectory

Shows how to order cells along a developmental or transitional path: choosing between StaVia and Slingshot, setting the grouping and starting point, and viewing the result.

1Open the Trajectory tab

Open Trajectory in the bottom drawer. A trajectory orders cells along a path, so you can follow how expression changes from one state to another.

Trajectory, step 1: Open the Trajectory tab

2Name the trajectory

Name the trajectory. The name starts as a template: {method} and {grouping} are replaced with your choices when you submit, so you can leave it alone or type something that describes the path, such as Neural lineage.

3Choose a method

StaVia and Slingshot both infer a path through your cells. StaVia is the default and the right choice for spatial data, and for anything large: it has no size limit. Slingshot suits smaller non-spatial datasets and cannot run on an embedding of more than 250,000 cells.

Trajectory, step 3: Choose a method

Note: if you choose Slingshot and then an embedding above 250,000 cells, iSCanGuide says so and asks you to pick a smaller embedding or switch to StaVia. The limit is enforced when the job runs as well, so a too-large embedding is refused rather than quietly cut down.

4Choose an embedding and grouping

Select the embedding to build the trajectory on, then how the cells are grouped. StaVia groups by clustering. Slingshot can also group by annotation or metadata.

Trajectory, step 4: Choose an embedding and grouping

5Choose a starting point

Tell iSCanGuide where the path begins by choosing the group that represents the earliest state. There is no default, and your choice sets the direction of the whole path. With Slingshot you can also name the groups the path should end at.

Trajectory, step 5: Choose a starting point

Which field appears

MethodFields
StaViaStart cluster
SlingshotStart groups and End groups

Both are required and neither has a default.

6Distance method

The distance method sets how similarity between groups is measured when the path is built. Slingshot is the default.

Trajectory, step 6: Distance method

Distance methods

Slingshot, Simple, Mutual nearest neighbors, Scaled Full, Scaled Diagonal.

7Advanced parameters

The advanced section holds the fitting parameters. The defaults suit most datasets.

Trajectory, step 7: Advanced parameters

Advanced parameters

FieldDefault
Convergence threshold0.001
Approximate number of points150
Stretch2
Allow breaksoff

8Existing trajectories

Click Run Analysis to start. Completed trajectories are listed with the method and embedding they used. Expand one to see its parameters.

Trajectory, step 8: Existing trajectories

9Visualize the trajectory

Create a view and choose your trajectory to colour cells by their position along the path.

Trajectory, step 9: Visualize the trajectory