GuangyuWangLab2021 / cellDancer

Predict RNA velocity through deep learning
https://guangyuwanglab2021.github.io/cellDancer_website/
BSD 3-Clause "New" or "Revised" License
60 stars 11 forks source link

The velocity1 and velocity2 values in celldancer_df are NaN #21

Open zcaiwei opened 8 months ago

zcaiwei commented 8 months ago

Hi, When I'm running the command cellDancer_df=compute_cell_velocity(cellDancer_df=cellDancer_df, projection_neighbor_choice='gene', expression_scale='power10', projection_neighbor_size=10, speed_up=(100,100)) , I'm having the following problem: NaN_plot Can you tell me how to tune the parameter to avoid generating NaN values?

Abclisy commented 8 months ago

Hello zcaiwei, thank you for reaching out!

If you would like to have the velocity1 and velocity2 for all cells, speed_up parameter could be set to None. This will consume more time to calculate the cell velocity. This parameter is used to downsample from the existing cells for the calculation of cell velocity. If you have any other questions, please feel free to contact us!

speed_up (optional, tuple (default: (60,60))) – Speed up by giving the sampling grid to downsample cells. None if all cells are used to compute cell velocity.