FlowJo now supports the import of CLR (CLassification Results) and CSV (Comma Separated Values) files. This document describes the various file formats, the results of importing them for a sample population, and how to use them in FlowJo.
These features are still early in their adoption and development.
Potential Use Case(s)
1. Doing a probabilistic study of many gates on the same set of events. Researcher wants to look at the mean gate of a set of gates.
1. File formats
The candidate release specification for CLR files is enumerated in its entirety here. It is valid CSV file, where each row corresponds to an event in the sample file. Each column corresponds to a ‘class’, where the value is the probability of the event being a member of the class, i.e. a value between 0 and 1. The first line may optionally contain class names. If the probability values are either 0 or 1, then the the CLR file is said to have ‘definite’ class assignments.
Here is an example of a general CLR file:
A B C
.1 .8 .1
.8 .1 .1
.6 .3 .1
.3 .2 .5
Here is an example of a definite CLR file:
A B C
0 1 0
1 0 0
1 0 0
0 0 1
The CSV format is similar to the CLR format, with the exception that the values are positive integers. The integer value corresponds to the class of which the event is a member.
Here is an example of a CSV file:
Note: the three examples above all define the same assignment of events to populations.
2. Importing CLR, CSV files
When a CLR or CSV file is imported for a sample, the following happens:
– For each column in the file, a new derived parameter is created. The rows in the file contain the data values for each parameter. To aid in visualization, the values are normalized to a data range of 4096.
– One or more range gates are created that contain all the events with a particular value. The gates are determined by the peak finder algorithm. For a definite CLR file, only a single gate is created on the included events.
3. How to use CLR and CSV files in FlowJo
Currently, the import of CLR or CSV files is only supported through the drag-n-drop mechanism. The user must drag an individual CLR/CSV file to the appropriate sample in the workspace, and the number of data rows in the CLR/CSV file must be equal to the number of events in the sample file.