$ cryodrgn downsample -h$ cryodrgn train_vae -h$ cryodrgn analyze -h$ cryodrgn eval_vol -h }}){kind=logo}
{{ a.cli[0] }}
{% elif a.w == 'number' %}
{% else %}
{% endif %}
{% endmacro %}
{% macro render_dataset_fieldset(cmd_key, datadir_id) %}
{% set ds = (command_builder_schema[cmd_key] | selectattr('title', 'equalto', 'Dataset loading') | list | first) %}
{% endmacro %}
Pick a primary reconstruction command. The line below lists required arguments plus options you actually set (empty fields and CLI defaults are omitted).
cryodrgn abinit …
This page does not run training — copy the line into your shell or job script.
 }})
Plot dataset particles by model outputs such as z-latent-space dimensions or analysis covariates such as UMAP embeddings.
Use lasso selection to write --ind .pkl files,
and optionally view dataset images or volume reconstructions for each particle.
 }})
Grid of scatters across latent dimensions and embeddings (PCA / UMAP on the diagonal when available).
{% endif %} {% if latent3d_active %} }})
Interactive 3D scatter of three latent axes, with optional color from analysis columns when you have at least three z dimensions.
 }})
Compose a cryodrgn command you can copy and paste for tools such as abinit and train_vae — required args plus only the options you fill in.
 }})
Draw a line through the latent scatter to generate volumes along a trajectory — direct interpolation or nearest-neighbor mode — with ChimeraX previews.
{% endif %} {# Inactive interfaces start on next row #} {% if any_inactive %} {% endif %} {% if not explorer_active %}
Plot dataset particles by model outputs such as z-latent-space dimensions or analysis covariates such as UMAP embeddings.
Use lasso selection to write --ind .pkl files.
{% if command_builder_only %}
Activate this interface by running cryodrgn dashboard <outdir>.
Activate this interface by loading an output directory with single-particle image previews.
{% endif %}
Grid of scatters across latent dimensions and embeddings (PCA / UMAP on the diagonal when available).
Activate this interface by running cryodrgn dashboard <outdir>.
Interactive 3D scatter of three latent axes, with optional color from analysis columns when you have at least three z dimensions.
Activate this interface by running cryodrgn dashboard <outdir>.
Activate this interface by loading a run with at least three latent dimensions (current: {{ zdim }}).
{% endif %}
Draw a line through the latent scatter to generate volumes along a trajectory — direct interpolation or nearest-neighbor mode — with ChimeraX previews.
{% if command_builder_only %}Activate this interface by running cryodrgn dashboard <outdir>.
Activate this interface by using a CUDA-enabled machine and loading a run with model weights for the selected epoch.
{% endif %}Each point is a particle in the space of three latent coordinates you choose. Optional color encodes another quantity from the analysis table.
{{ reason }}
Each panel compares two latent coordinates. The diagonal shows a 2D embedding; one triangle colors particles by a covariate; the other shows a light scatter or density pattern.
The pairwise grid needs at least two latent coordinates. This reconstruction has {{ n }}.
{% elif kind == "z3" %}The 3-D Latent Space Visualizer needs at least three latent coordinates. This reconstruction has {{ n }}.
{% elif kind == "pca" %}The diagonal embedding needs PCA components in the analysis table for this epoch.
{% else %}There is no numeric column available to color the lower panels.
{% endif %}Choose covariates for plot axes and point coloring. Use lasso selection to create a particle filtering index; generate images and volumes with or without a selection using the interface on the right.