The FlyWire connectome was generated using a combination of automated image analysis and human proofreading. First, a team of researchers used electron microscopy to image a complete adult fly brain (FAFB). The resulting images were then aligned and segmented using AI and other algorithms. This resulted in a preliminary connectome, that was then proofread by a community of scientists and citizen scientists using the FlyWire proofreading platform. The platform allows users to view and edit the connectome, and to provide annotations about the neurons and synapses. This process helped to improve the accuracy of the connectome, and to identify new features that were not visible in the original images. For additional information refer to the FlyWire homepage.

In FlyWire, neuron side is decided based on SOMA (cell body) position, and both neuropil and neuron side annotations are from fly's perspective. After the initial FlyWire release, we learned that the FAFB dataset used in FlyWire was left/right inverted during image acquisition. This impacted neurons labeled L/R, neuropil names, and any other use of left/right. This issue was addressed by inverting the left/right orientation of all annotations and labels in both CAVE and Codex. However, we decided not to change the imagery or segmentation data, which means that all images and 3D views are inverted, both for cells and neuropils. If you have any questions, please contact flywire@princeton.edu

Codex provides access to static FlyWire connectome data releases (sometimes referred to as 'snapshots' or 'versions'). These releases include only cells that were marked as proofread. In addition, Codex excludes weak connections. Specifically, a pair of cells counts as connected if the number of synapses is at least 5. To switch from the default data snapshot one of the older releases use the dataset dropdown menu. And for live queries against the proofreading database tables use the CAVE Client Python library instead.

The connectome forms a directed graph where nodes correspond to neurons and edges to synaptic connections. Edges have the following properties: number of synapses, region (neuropil), predicted neurotransmitter type. Nodes have various annotations, including hierarchical classification, identification labels, neurotransmitter predictions, side, and more. A pair of nodes can be connected with multiple edges corresponding to different brain regions.

Yes, visit the data download portal to review and download the FlyWire connectome data resources.

For individual cell identification data (e.g. cell type annotations) you can navigate to the cell detail page and submit the annotations directly from there. For bulk deposit, please email a CSV (or other plain text file format) where cell id / annotation columns are clearly marked to arie@princeton.edu. In the email include a description of the data + citation / source.

Yes, for certain attributes (root_id, label, side, nt_type, input_neuropils, output_neuropils, input_hemisphere, output_hemisphere, flow, super_class, class, sub_class, cell_type, resolved_type, gene, hemilineage, nerve, name, group, connectivity_tag, mirror_twin_root_id, marker) you can search using structured query terms e.g. by typing into search box class == JON or nt_type != GABA. You can also chain the terms with and/or rules like so: class == JON && nt_type != GABA or similarly class == JON || class == olfactory || class == dsx etc. (but different chaining rules cannot be mixed).
Search attributes

• root_id ID of the cell, unique across data versions but might get replaced if altered by proofreading
• label Human readable label assigned during cell identification process (each cell can have zero or more labels)
• side Neuron side / hemisphere [left, right, center]
• nt_type Neuro-transmitter type [ACH, DA, GABA, GLUT, OCT, SER]
• input_neuropils Brain region / neuropil with upstream synaptic connections [AL_L, AL_R, AME_L, AME_R, AMMC_L, AMMC_R, AOTU_L, AOTU_R, ATL_L, ATL_R, AVLP_L, AVLP_R, BU_L, BU_R, CAN_L, CAN_R, CRE_L, CRE_R, EB, EPA_L, EPA_R, FB, FLA_L, FLA_R, GA_L, GA_R, GNG, GOR_L, GOR_R, IB_L, IB_R, ICL_L, ICL_R, IPS_L, IPS_R, LAL_L, LAL_R, LA_L, LA_R, LH_L, LH_R, LOP_L, LOP_R, LO_L, LO_R, MB_CA_L, MB_CA_R, MB_ML_L, MB_ML_R, MB_PED_L, MB_PED_R, MB_VL_L, MB_VL_R, ME_L, ME_R, NO, OCG, PB, PLP_L, PLP_R, PRW, PVLP_L, PVLP_R, SAD, SCL_L, SCL_R, SIP_L, SIP_R, SLP_L, SLP_R, SMP_L, SMP_R, SPS_L, SPS_R, UNASGD, VES_L, VES_R, WED_L, WED_R]
• output_neuropils Brain region / neuropil with downstream synaptic connections [AL_L, AL_R, AME_L, AME_R, AMMC_L, AMMC_R, AOTU_L, AOTU_R, ATL_L, ATL_R, AVLP_L, AVLP_R, BU_L, BU_R, CAN_L, CAN_R, CRE_L, CRE_R, EB, EPA_L, EPA_R, FB, FLA_L, FLA_R, GA_L, GA_R, GNG, GOR_L, GOR_R, IB_L, IB_R, ICL_L, ICL_R, IPS_L, IPS_R, LAL_L, LAL_R, LA_L, LA_R, LH_L, LH_R, LOP_L, LOP_R, LO_L, LO_R, MB_CA_L, MB_CA_R, MB_ML_L, MB_ML_R, MB_PED_L, MB_PED_R, MB_VL_L, MB_VL_R, ME_L, ME_R, NO, OCG, PB, PLP_L, PLP_R, PRW, PVLP_L, PVLP_R, SAD, SCL_L, SCL_R, SIP_L, SIP_R, SLP_L, SLP_R, SMP_L, SMP_R, SPS_L, SPS_R, UNASGD, VES_L, VES_R, WED_L, WED_R]
• input_hemisphere Brain hemisphere / side with upstream synaptic connections [Center, Left, Right]
• output_hemisphere Brain hemisphere / side with downstream synaptic connections [Center, Left, Right]
• flow Flow, refers to containment of the neuron in the brain [intrinsic, efferent, afferent]
• super_class Cell typing attribute, indicates function or other property of the cell [optic, central, sensory, visual_projection, ascending, descending, visual_centrifugal, motor, endocrine]
• class Cell typing attribute, indicates function or other property of the cell [ALIN, ALLN, ALON, ALPN, AN, CX, DAN, Kenyon_Cell, LHCENT, LHLN, MBIN, MBON, TPN, TuBu, bilateral, gustatory, hygrosensory, mechanosensory, ocellar, olfactory, optic_lobe_intrinsic, optic_lobes, pars_intercerebralis, pars_lateralis, thermosensory, unknown_sensory, visual]
• sub_class Cell typing attribute, indicates function or other property of the cell
• cell_type Cell typing attribute, indicates function or other property of the cell
• resolved_type Primary cell type (if more than one are assigned).
• gene Gene expression [Doublesex, Fruitless]
• hemilineage Lineage from Janelia hemibrain dataset
• nerve Nerve type, if applicable [CV, AN, MxLbN, OCN, PhN, aPhN, NCC, ON]
• name Automatically assigned name (based on most significant input/output regions) - unique across data versions, but might get replaced if affected by proofreading
• group Automatically assigned group name (based on most significant input/output regions)
• connectivity_tag Connectivity tag, describes the type of connections the neuron participates in [3_cycle_participant, broadcaster, feedforward_loop_participant, highly_reciprocal_neuron, integrator, nsrn, reciprocal, rich_club]
• mirror_twin_root_id ID of the mirror-twin cell, optional
• marker Generic cell markers

Binary operators

• {equal} (short alternative ==) Binary, LHS attribute of the cell and RHS value are equal
• {not_equal} (short alternative !=) Binary, LHS attribute of the cell and RHS value are not equal
• {starts_with} (short alternative ^*) Binary, LHS attribute of the cell starts with RHS value (e.g., label {starts_with} LC)
• {contains} (short alternative >>) Binary, LHS attribute of the cell contains RHS value (e.g., label {contains} dsx)
• {not_contains} (short alternative !>) Binary, LHS attribute of the cell does not contain RHS value (e.g., label {not_contains} dsx)
• {in} (short alternative <<) Binary, LHS attribute of the cell equals one of the comma-separated values on RHS
• {not_in} (short alternative !<) Binary, LHS attribute of the cell is not equal to any of the comma-separated values on RHS
• {upstream_region} (short alternative ^R) Binary, matches cells upstream of RHS, with synapses in LHS region, where region is either hemisphere (left/right/center) or neuropil (e.g. GNG)
• {downstream_region} (short alternative !R) Binary, matches cells downstream of RHS, with synapses in LHS region, where region is either hemisphere (left/right/center) or neuropil (e.g. GNG)
• {pathways} (short alternative =>) Binary, match all cells along shortest-path pathways from LHS to RHS

Unary operators

• {has} (short alternative $$) Unary, attribute of the cell has value (not empty)
• {missing} (short alternative !$) Unary, attribute of the cell has no value (empty)
• {upstream} (short alternative ^^) Unary, matches cells upstream of specified Cell ID
• {downstream} (short alternative !^) Unary, matches cells downstream of specified Cell ID
• {reciprocal} (short alternative ^v) Unary, matches reciprocal-feedback cells that are both downstream and upstream of specified Cell ID
• {similar_shape} (short alternative ~~) Unary, matches cells that are similar in shape to specified Cell ID
• {similar_connectivity_upstream} (short alternative ~u) Unary, matches cells that have similar upstream connectivity to specified Cell ID
• {similar_connectivity_downstream} (short alternative ~d) Unary, matches cells that have similar downstream connectivity to specified Cell ID
• {similar_connectivity} (short alternative ~c) Unary, matches cells that have similar connectivity (both up and downstream) to specified Cell ID
• {similar_projection_upstream} (short alternative ~pu) Unary, matches cells that have similar input type projections to specified Cell ID
• {similar_projection_downstream} (short alternative ~pd) Unary, matches cells that have similar output type projections to specified Cell ID
• {similar_projection} (short alternative ~pc) Unary, matches cells that have similar input/output type projections to specified Cell ID

Logical chaining operators

• {upstream_intersection} (short alternative ~UI) Unary, matches cells upstream of ALL matching cells on the RHS
• {downstream_intersection} (short alternative ~DI) Unary, matches cells downstream of ALL matching cells on the RHS
• {upstream_union} (short alternative ~UU) Unary, matches cells upstream of ANY of the matching cells on the RHS
• {downstream_union} (short alternative ~DU) Unary, matches cells downstream of ANY of the matching cells on the RHS
• {and} (short alternative &&) N-ary, logical AND / intersection of the operands
• {or} (short alternative ||) N-ary, logical OR / union of the operands
• {not} (short alternative !!) Unary, logical NOT / complement of the operand on the RHS

Logical operators can be combined and prioritized using double curly braces {{ and }}.See sample queries for specific examples.

Cell names were assigned automatically based on neuropils (brain regions) where the cells have most of their synapstic connections. Note of caution: this naming scheme will be carried forward with next releases, so as proofreading continues names of certain cells might change over time.

Sensory, ascending, descending and visual projection neurons were identified by the Jefferis Lab. The remaining neurons were assigned to the central brain if they had more than 100 connections in the central brain (vs the optic lobe), and to the optic lobe otherwise. As cell identification efforts progress, these classes will evolve too over time.

A minimum threshold of 5 synapses is applied across the board to determine connectivity between pairs of cells.

NBLAST is a method for assigning similarity scores to pairs of neurons. NBLAST scores are not necessarily symmetric, and they range from 1 (morphologically similar) to -1 (not similar).
The NBLAST scores used here were contributed by Dr. Philipp Schlegel.

Shortest path length is equivalent to the minimum number of 'hops' required to reach from the source cell to the target via synaptic connections in the connectome network.
NOTE Only connections with 5+ synapses are taken into account.

The Neuroglancer based 3D viewer generally uses coordinate system with 4, 4, 40 nm voxels (shown in the upper left), while outside of the viewer points (e.g. synapse locations) are generally specified in nanometer resolution. The origin of the 3D vieer was chosen as a corner of the bounding box surrounding the image volume. The orientation of the coordinate system is roughly as follows: the x-axis runs from left to right (lateral), the y-axis from dorsal to ventral, and the z-axis from anterior to posterior. These are rough alignments because they were put in place when the sample was cut.

You can request to join by using this invitation link, or emailing flywire@princeton.edu (please include your name and affiliation).