Bridge a Deng et al. 2019 LSM volume into BANC voxel coordinates

Source: R/lm_pipeline.R

Top-level orchestrator for the Deng pipeline. Chains Stage 0 (LSM extract) + Stage A' (Elastix native -> JRC2018U) + Stage B” (H5 JRC2018U -> JRC2018F) + the existing Stage C (transformix JRC2018F -> BANC) + Stage D (nrrd_to_precomputed), producing a Neuroglancer precomputed image layer named <gene>_<sample>.ng/ on the BANC grid.

lsm_to_banc_layer(
  input,
  gene,
  sample,
  channel = "GFP",
  dataset = "deng_et_al_2019",
  output_dir,
  source_path = NULL,
  chunk_size = c(64L, 64L, 64L),
  keep_intermediates = FALSE,
  threads = 8L,
  verbose = TRUE
)

Arguments

input

source .lsm path.

gene

short gene name; forms layer name <gene>_<sample>.

sample

sample identifier (e.g. BRAIN-F, VNC-M, or any user-chosen short tag).

channel

the upstream channel string recorded in the registry (default "GFP").

dataset

master folder under light_level/ (default "deng_et_al_2019").

output_dir

parent directory for intermediates + final precomputed layer.

source_path

optional original storage path for provenance.

chunk_size

length-3 integer chunk size for the precomputed layer.

keep_intermediates

logical; keep the per-stage intermediate NRRDs (default FALSE).

threads

passed to Elastix + RenderTransformed.

Value

list with precomputed_dir, registry_entry (one-row tibble), timings (named numeric, s).

Software prerequisites

  • Elastix 5.x elastix + transformix on PATH

  • Java 17+ + the saalfeldlab RenderTransformed JAR

  • Python (via reticulate) with tifffile + SimpleITK

  • nat.jrcbrains H5 registrations downloaded (nat.jrcbrains::download_saalfeldlab_registrations())

  • JRC2018U template NRRD at ~/templates/JRC2018_UNISEX_20x_HR.nrrd (download from janelia-flylight-templates S3).