NeuroM currently supports the SWC format, the BBP HDF5 formats, and offers experimental support for NeuroLucida .asc files.
The morphology definitions page<definitions-label> for definitions of concepts such as point<point-label>, section<section-label>, soma<soma-label> and neurite<neurite-label> in NeuroM.
Complete this section with additional NeuroM specific restrictions on the formats below.
The SWC format represents a neuron as a set of trees that are connected to a soma. The neuronal morphology is encoded as a rooted tree of 3D points and the corresponding radii. More information can be found here.
The soma format is considered to be a series of connected cylinders if there are 2 or more points, and a sphere if there is only a single point.
In the special case of a 3 point soma where points follow the following scheme:
| 1 | 2 | 4 | 5 | 6 | 7 | 8 |
|---|---|---|---|---|---|---|
|
Type |
|
|
Z |
|
Parent ID |
| === | ==== | == | ======= | == | ======= | ========= |
|
|
xs |
|
zs |
|
|
|
|
xs | (ys-rs) | zs |
|
|
|
|
xs | (ys+rs) | zs |
|
|
They are detected as the as NeuroMorpho SWC soma style, and used that way: see :pySoma<neurom.core._soma.SomaThreePointCylinders> for more details.
It is not considered a good practice to represent the same section of the tree in different places within a file, but it is not forbidden. The parent ID should always be smaller that the current ID of a point.
Add reference to SWC paper and more semantic constraints.
Add semantic constraints on different soma types once these have been determined. For more info on what is to be considered, see neuromorpho.org's.
The HDF5 morphology formats developed by the BBP represent the neuron as a tree of sections<section-label>. The specifications for the two versions of the format can be found in the HBP morphology format documentation page.
The NeuroLucida .asc file format is commonly used but lacking in an open format specification. NeuroM provides a best-effort experimental reader that parses information equivalent to the two formats above, that is to say, it does not deal with annotations or other meta-data, and is restricted purely to the topological and geometrical features of a neuron, as well as the neurite type information.
Warning
The NeuroLucida parser is experimental. Use at own risk when extracting numerical information. We make no statement as to the correctness of numerical output.
References and more information?