Read BANC CAVE-tables, good sources of metadata

CAVE tables query functions that track neurons across segmentation changes so that annotations and neuron entities can be stably tracked together. The Brain And Nerve Cord (BANC) dataset represents the first complete connectome including both brain and ventral nerve cord of a limbed animal, comprising approximately 160,000 neurons across the entire central nervous system.

Usage

banc_cave_tables(datastack_name = NULL, select = NULL)

banc_cave_views(datastack_name = NULL, select = NULL)

banc_edgelist(
  version = c("v2", "v3"),
  source = c("gcs", "cave"),
  overwrite = FALSE,
  edgelist_view = NULL,
  ...
)

banc_mitochondria(
  rootids = NULL,
  table = "mitochondria_v1",
  rawcoords = FALSE,
  chunk_size = 200000L,
  ...
)

banc_nuclei(
  rootids = NULL,
  nucleus_ids = NULL,
  table = c("both", "somas_v1a", "somas_v1b"),
  rawcoords = FALSE,
  source = c("gcs", "cave"),
  fallback = TRUE,
  ...
)

banc_cell_info(
  rootids = NULL,
  rawcoords = FALSE,
  source = c("gcs", "cave"),
  fallback = TRUE,
  ...
)

banc_proofreading_notes(
  rootids = NULL,
  rawcoords = FALSE,
  source = c("gcs", "cave"),
  fallback = TRUE,
  ...
)

banc_cell_ids(rootids = NULL, ...)

banc_neck_connective_neurons(
  rootids = NULL,
  table = c("neck_connective_y92500", "neck_connective_y121000"),
  source = c("gcs", "cave"),
  fallback = TRUE,
  ...
)

banc_peripheral_nerves(
  rootids = NULL,
  source = c("gcs", "cave"),
  fallback = TRUE,
  ...
)

banc_backbone_proofread(
  rootids = NULL,
  source = c("gcs", "cave"),
  fallback = TRUE,
  ...
)

banc_nt_prediction(
  rootids = NULL,
  table = "synapses_250226_nt_prediction_5",
  simplify = TRUE,
  rawcoords = TRUE,
  ...
)

banc_version()

Arguments

datastack_name: Defaults to "brain_and_nerve_cord". See https://global.daf-apis.com/info/ for other options.
select: A regex term for the name of the table you want
version: Character, "v2" (default, paper synapses) or "v3" (updated synapses, still in testing).
source: "gcs" (default; reads the public compiled feather) or "cave" (live materialised view query).
overwrite: Logical. If TRUE and source = "gcs", re-download the cached feather.
edgelist_view: Optional CAVE view name override (only honoured when source = "cave"). Defaults are derived from version: synapses_v<version>_backbone_proofread_and_peripheral_nerves_counts.
...: Additional arguments passed to flywire_cave_query or bancr:::get_cave_table_data.
rootids: Character vector specifying one or more BANC rootids. As a convenience this argument is passed to banc_ids allowing you to pass in data.frames, BANC URLs or simple ids.
table: Character, possible alternative tables for the sort of data frame the function returns. One must be chosen.
rawcoords: Logical, whether or not to convert from raw coordinates into nanometers. Default is FALSE.
chunk_size: Integer page size for full-table pulls (used only when rootids = NULL). The mitochondria_v1 table has millions of rows and a single materialised response trips reticulate's string parser (Error: basic_string); paginating by limit/offset keeps each response small enough to cross the R/Python boundary. Default 200000.
nucleus_ids: Character vector specifying one or more BANC nucleus ids. The nucleus (https://en.wikipedia.org/wiki/Cell_nucleus) contains the cell body and provides a stable reference point for neuron identification.
fallback: Logical, default TRUE. If the primary source errors, automatically retry against the alternative source and emit a warning describing what happened. Set FALSE to surface the original error directly.
simplify: logical, if TRUE then the proportion of presynaptic connections for each transmitter type is returned, for each query neuron.

Value

A data.frame describing a CAVE-table related to the BANC project. In the case of banc_cave_tables, a vector is returned containing the names of all query-able cave tables.

Details

CAVE tables store rich metadata supporting analysis of distributed neural control across the entire central nervous system. For more information about CAVE infrastructure, see https://www.caveconnecto.me/CAVEclient/.

banc_edgelist returns a data frame of neuron-neuron connections where the pre (presynaptic) neuron is upstream of the post (postsynaptic) neuron. This edgelist contains synaptic connectivity data crucial for understanding distributed neural control and behaviour-centric neural modules across the brain-VNC boundary.

Two synapse-table versions are exposed via the version argument: "v2" (default) is the paper-version edgelist built from CAVE synapses_v2; "v3" is the updated/refined synapse table (CAVE synapses_v3, still in testing — see banc_all_synapses()). The two sources differ slightly in coverage; for most analyses "v3" is the closer-to-current snapshot and "v2" matches the published numbers.

Two backing stores are supported via the source argument: "gcs" (default) reads the pre-computed compiled feather at gs://lee-lab_brain-and-nerve-cord-fly-connectome/compiled_data/banc_888/banc_888_edgelist_simple_<version>.feather from the public bucket — no BANC authentication needed, ~285 MB download for v2 / ~336 MB for v3, cached locally under tools::R_user_dir("bancr", "cache"). The returned schema is pre, post, count, norm, post_count, pre_count.

source = "cave" runs a live materialised CAVE view query (the previous default). Requires authenticated CAVE access; the returned schema includes pre_pt_root_id, post_pt_root_id and n. Use this when you need labels fresher than the latest GCS snapshot or want to override the materialisation timestamp via ....

banc_cell_info accesses the cell_info CAVE table containing non-centralised annotations from the research community for connectome neurones. These annotations represent diverse contributions from researchers studying specific neural circuits and cell types in the BANC dataset.

Examples