MultiMM 1.0.5__tar.gz → 1.0.7__tar.gz

{multimm-1.0.5 → multimm-1.0.7}/MultiMM.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: MultiMM
-Version: 1.0.5
+Version: 1.0.7
 Summary: A tool for chromatin modeling from nucleosomes to chromosomal territories.
 Home-page: https://github.com/SFGLab/MultiMM
 Author: Sebastian Korsak
@@ -30,7 +30,7 @@ Requires-Dist: OpenMM
 
 # MultiMM: An OpenMM-based software for whole genome 3D structure reconstruction
 
-MultiMM is an OpenMM model designed for modeling the 3D structure of the whole genome. Its distinguishing feature is that it is multiscale, meaning it aims to model different levels of chromatin organization, from smaller scales (nucleosomes) to the level of chromosomal territories. The algorithm is both fast and accurate. A key feature enabling its speed is GPU parallelization via OpenMM, along with smart assumptions that assist the optimizer in finding the global minimum. One such fundamental assumption is the use of a Hilbert curve as the initial structure. This helps MultiMM to converge faster because the initial structure is already highly compacted.
+MultiMM is an OpenMM model designed for modeling the 3D structure of the whole *human* genome. Its distinguishing feature is that it is multiscale, meaning it aims to model different levels of chromatin organization, from smaller scales (nucleosomes) to the level of chromosomal territories. The algorithm is both fast and accurate. A key feature enabling its speed is GPU parallelization via OpenMM, along with smart assumptions that assist the optimizer in finding the global minimum. One such fundamental assumption is the use of a Hilbert curve as the initial structure. This helps MultiMM to converge faster because the initial structure is already highly compacted.
 
 ![GW_em](https://github.com/user-attachments/assets/3d019616-2c0f-4bfc-a792-d87fcdc0d96a)
 
@@ -40,23 +40,25 @@ After running the MultiMM model, users obtain a genome-wide structure. Users can
 
 ## Key Features
 
+- OpenMM based.
+- User-friendly software available via PyPI.
 - Efficient simulation of chromatin interactions using 3D conformations.
-- Can simulate the scales of nucleosomes, TADs, compartments, chromosomal territories and interactions with lamina.
-- Integration with **bedpe** for chromatin structures.
-- Scalable simulations across different force fields and resolution levels.
-- Compatible with modern GPU-accelerated simulation libraries.
+- Can simulate the scales of nucleosomes, TADs, compartments, chromosomal territories and interactions with lamina. Scalable simulations across different force fields and resolution levels.
+- Compatible with modern GPU-accelerated simulation libraries. CPU acceleration can also be done.
 
 ## About Operating Systems
 
 MultiMM has been tested primarily on Linux-based operating systems, with successful tests in Ubuntu, Debian, and Red Hat-based distributions. It is also possible to run it on macOS, though without CUDA support, which is helpful for accelerating computations. We do not recommend running MultiMM on Windows systems.
 
 ## Installation
-Create a Python 3.10 environment and type:
+MultiMM can be easily installed with pip:
 
 ```bash
-pip install -r requirements.txt
+pip install MultiMM
 ```
 
+PyPI software: https://pypi.org/project/MultiMM/.
+
 ## Input Data
 
 MultiMM relies on three types of datasets:
@@ -128,7 +130,7 @@ TRJ_FRAMES = 100
 After specifying the parameters and forces, users can run the following command in the terminal:
 
 ```bash
-python run.py -c config.ini
+MultiMM -c config.ini
 ```
 
 The software will output a folder with the resulting structure and plots showing compartment distribution.
@@ -137,6 +139,28 @@ Example data can be found on Google Drive: https://drive.google.com/drive/folder
 
 In the `examples` folder, we provide example configuration files for different modeling scenarios.
 
+## Vizualization
+
+For vizualization purposes, if you would like to import the whole genome structure you may use the command,
+
+```python
+import simulation.plots as splt
+
+splt.viz_chroms(sim_path)
+```
+
+For `sim_path` you should add the output folder directory path (add `comps=Fase` in case that you do not need compartment coloring). Otherwise, in case that the user would like to model a particular region, without using compartment or chromosome coloring (pretty much any cif structure), they can type,
+
+```python
+import simulation.plots as splt
+import simulation.utils as suts
+
+V = suts.get_coordinates_cif(cif_path)
+splt.viz_structure(V)
+``` 
+
+We would like once again to thank people who developed `pyvista` library and allow us to have fast and good vizualizations of large chromatin structure.
+
 ## Simulation Arguments
 
 MultiMM has numerous configurable parameters. Below is a description of each argument and its default values. The defaults have been tested for genome-wide simulation but can be modified in the configuration file if needed. Units are typically assumed based on OpenMM conventions, though explicit unit specification is not required.
@@ -151,7 +175,7 @@ This version should now be more precise and polished. Let me know if you'd like
 | DEVICE                       | str          | None        | None          | device index for CUDA or OpenCL (count from 0) |
 | INITIAL_STRUCTURE_PATH       | str          | None        | None          | Path to CIF file. |
 | BUILD_INITIAL_STRUCTURE      | bool         | True        | None          | To build a new initial structure. |
-| INITIAL_STRUCTURE_TYPE       | str          | hilbert     | None          | you can choose between: hilbert, circle. |
+| INITIAL_STRUCTURE_TYPE       | str          | hilbert     | None          | you can choose between: hilbert, circle, rw, confined_rw, self_avoiding_rw, helix, spiral, sphere, knot. |
 | FORCEFIELD_PATH              | str          | None        | None          | Path to XML file with forcefield. |
 | N_BEADS                      | int          | 50000       | None          | Number of Simulation Beads. |
 | COMPARTMENT_PATH             | str          | None        | None          | It should be  a `.bed` file with subcompartments from Calder. |
@@ -200,7 +224,6 @@ This version should now be more precise and polished. Let me know if you'd like
 | CF_USE_CENTRAL_FORCE         | bool         | False       | None          | Attraction of smaller chromosomes to the nucleolus (in the center). |
 | CF_STRENGTH                  | float        | 10.0       | kJ/mol        | Strength of Attraction of Smaller Chromosomes |
 | NUC_DO_INTERPOLATION         | bool         | False       | None          | Attraction of smaller chromosomes. |
-| MAX_NUCS_PER_BEAD            | int          | 4           | None          | Maximum amount of nucleosomes per single bead for nucleosome interpolation. |
 | NUC_RADIUS                   | float        | 0.1         | None          | The radius of the single nucleosome helix. |
 | POINTS_PER_NUC               | int          | 20          | None          | The number of points that consist a nucleosome helix. |
 | PHI_NORM                     | float        | pi/5        | None          | Zig zag angle. |

{multimm-1.0.5 → multimm-1.0.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: MultiMM
-Version: 1.0.5
+Version: 1.0.7
 Summary: A tool for chromatin modeling from nucleosomes to chromosomal territories.
 Home-page: https://github.com/SFGLab/MultiMM
 Author: Sebastian Korsak
@@ -30,7 +30,7 @@ Requires-Dist: OpenMM
 
 # MultiMM: An OpenMM-based software for whole genome 3D structure reconstruction
 
-MultiMM is an OpenMM model designed for modeling the 3D structure of the whole genome. Its distinguishing feature is that it is multiscale, meaning it aims to model different levels of chromatin organization, from smaller scales (nucleosomes) to the level of chromosomal territories. The algorithm is both fast and accurate. A key feature enabling its speed is GPU parallelization via OpenMM, along with smart assumptions that assist the optimizer in finding the global minimum. One such fundamental assumption is the use of a Hilbert curve as the initial structure. This helps MultiMM to converge faster because the initial structure is already highly compacted.
+MultiMM is an OpenMM model designed for modeling the 3D structure of the whole *human* genome. Its distinguishing feature is that it is multiscale, meaning it aims to model different levels of chromatin organization, from smaller scales (nucleosomes) to the level of chromosomal territories. The algorithm is both fast and accurate. A key feature enabling its speed is GPU parallelization via OpenMM, along with smart assumptions that assist the optimizer in finding the global minimum. One such fundamental assumption is the use of a Hilbert curve as the initial structure. This helps MultiMM to converge faster because the initial structure is already highly compacted.
 
 ![GW_em](https://github.com/user-attachments/assets/3d019616-2c0f-4bfc-a792-d87fcdc0d96a)
 
@@ -40,23 +40,25 @@ After running the MultiMM model, users obtain a genome-wide structure. Users can
 
 ## Key Features
 
+- OpenMM based.
+- User-friendly software available via PyPI.
 - Efficient simulation of chromatin interactions using 3D conformations.
-- Can simulate the scales of nucleosomes, TADs, compartments, chromosomal territories and interactions with lamina.
-- Integration with **bedpe** for chromatin structures.
-- Scalable simulations across different force fields and resolution levels.
-- Compatible with modern GPU-accelerated simulation libraries.
+- Can simulate the scales of nucleosomes, TADs, compartments, chromosomal territories and interactions with lamina. Scalable simulations across different force fields and resolution levels.
+- Compatible with modern GPU-accelerated simulation libraries. CPU acceleration can also be done.
 
 ## About Operating Systems
 
 MultiMM has been tested primarily on Linux-based operating systems, with successful tests in Ubuntu, Debian, and Red Hat-based distributions. It is also possible to run it on macOS, though without CUDA support, which is helpful for accelerating computations. We do not recommend running MultiMM on Windows systems.
 
 ## Installation
-Create a Python 3.10 environment and type:
+MultiMM can be easily installed with pip:
 
 ```bash
-pip install -r requirements.txt
+pip install MultiMM
 ```
 
+PyPI software: https://pypi.org/project/MultiMM/.
+
 ## Input Data
 
 MultiMM relies on three types of datasets:
@@ -128,7 +130,7 @@ TRJ_FRAMES = 100
 After specifying the parameters and forces, users can run the following command in the terminal:
 
 ```bash
-python run.py -c config.ini
+MultiMM -c config.ini
 ```
 
 The software will output a folder with the resulting structure and plots showing compartment distribution.
@@ -137,6 +139,28 @@ Example data can be found on Google Drive: https://drive.google.com/drive/folder
 
 In the `examples` folder, we provide example configuration files for different modeling scenarios.
 
+## Vizualization
+
+For vizualization purposes, if you would like to import the whole genome structure you may use the command,
+
+```python
+import simulation.plots as splt
+
+splt.viz_chroms(sim_path)
+```
+
+For `sim_path` you should add the output folder directory path (add `comps=Fase` in case that you do not need compartment coloring). Otherwise, in case that the user would like to model a particular region, without using compartment or chromosome coloring (pretty much any cif structure), they can type,
+
+```python
+import simulation.plots as splt
+import simulation.utils as suts
+
+V = suts.get_coordinates_cif(cif_path)
+splt.viz_structure(V)
+``` 
+
+We would like once again to thank people who developed `pyvista` library and allow us to have fast and good vizualizations of large chromatin structure.
+
 ## Simulation Arguments
 
 MultiMM has numerous configurable parameters. Below is a description of each argument and its default values. The defaults have been tested for genome-wide simulation but can be modified in the configuration file if needed. Units are typically assumed based on OpenMM conventions, though explicit unit specification is not required.
@@ -151,7 +175,7 @@ This version should now be more precise and polished. Let me know if you'd like
 | DEVICE                       | str          | None        | None          | device index for CUDA or OpenCL (count from 0) |
 | INITIAL_STRUCTURE_PATH       | str          | None        | None          | Path to CIF file. |
 | BUILD_INITIAL_STRUCTURE      | bool         | True        | None          | To build a new initial structure. |
-| INITIAL_STRUCTURE_TYPE       | str          | hilbert     | None          | you can choose between: hilbert, circle. |
+| INITIAL_STRUCTURE_TYPE       | str          | hilbert     | None          | you can choose between: hilbert, circle, rw, confined_rw, self_avoiding_rw, helix, spiral, sphere, knot. |
 | FORCEFIELD_PATH              | str          | None        | None          | Path to XML file with forcefield. |
 | N_BEADS                      | int          | 50000       | None          | Number of Simulation Beads. |
 | COMPARTMENT_PATH             | str          | None        | None          | It should be  a `.bed` file with subcompartments from Calder. |
@@ -200,7 +224,6 @@ This version should now be more precise and polished. Let me know if you'd like
 | CF_USE_CENTRAL_FORCE         | bool         | False       | None          | Attraction of smaller chromosomes to the nucleolus (in the center). |
 | CF_STRENGTH                  | float        | 10.0       | kJ/mol        | Strength of Attraction of Smaller Chromosomes |
 | NUC_DO_INTERPOLATION         | bool         | False       | None          | Attraction of smaller chromosomes. |
-| MAX_NUCS_PER_BEAD            | int          | 4           | None          | Maximum amount of nucleosomes per single bead for nucleosome interpolation. |
 | NUC_RADIUS                   | float        | 0.1         | None          | The radius of the single nucleosome helix. |
 | POINTS_PER_NUC               | int          | 20          | None          | The number of points that consist a nucleosome helix. |
 | PHI_NORM                     | float        | pi/5        | None          | Zig zag angle. |

{multimm-1.0.5 → multimm-1.0.7}/README.md

@@ -1,6 +1,6 @@
 # MultiMM: An OpenMM-based software for whole genome 3D structure reconstruction
 
-MultiMM is an OpenMM model designed for modeling the 3D structure of the whole genome. Its distinguishing feature is that it is multiscale, meaning it aims to model different levels of chromatin organization, from smaller scales (nucleosomes) to the level of chromosomal territories. The algorithm is both fast and accurate. A key feature enabling its speed is GPU parallelization via OpenMM, along with smart assumptions that assist the optimizer in finding the global minimum. One such fundamental assumption is the use of a Hilbert curve as the initial structure. This helps MultiMM to converge faster because the initial structure is already highly compacted.
+MultiMM is an OpenMM model designed for modeling the 3D structure of the whole *human* genome. Its distinguishing feature is that it is multiscale, meaning it aims to model different levels of chromatin organization, from smaller scales (nucleosomes) to the level of chromosomal territories. The algorithm is both fast and accurate. A key feature enabling its speed is GPU parallelization via OpenMM, along with smart assumptions that assist the optimizer in finding the global minimum. One such fundamental assumption is the use of a Hilbert curve as the initial structure. This helps MultiMM to converge faster because the initial structure is already highly compacted.
 
 ![GW_em](https://github.com/user-attachments/assets/3d019616-2c0f-4bfc-a792-d87fcdc0d96a)
 
@@ -10,23 +10,25 @@ After running the MultiMM model, users obtain a genome-wide structure. Users can
 
 ## Key Features
 
+- OpenMM based.
+- User-friendly software available via PyPI.
 - Efficient simulation of chromatin interactions using 3D conformations.
-- Can simulate the scales of nucleosomes, TADs, compartments, chromosomal territories and interactions with lamina.
-- Integration with **bedpe** for chromatin structures.
-- Scalable simulations across different force fields and resolution levels.
-- Compatible with modern GPU-accelerated simulation libraries.
+- Can simulate the scales of nucleosomes, TADs, compartments, chromosomal territories and interactions with lamina. Scalable simulations across different force fields and resolution levels.
+- Compatible with modern GPU-accelerated simulation libraries. CPU acceleration can also be done.
 
 ## About Operating Systems
 
 MultiMM has been tested primarily on Linux-based operating systems, with successful tests in Ubuntu, Debian, and Red Hat-based distributions. It is also possible to run it on macOS, though without CUDA support, which is helpful for accelerating computations. We do not recommend running MultiMM on Windows systems.
 
 ## Installation
-Create a Python 3.10 environment and type:
+MultiMM can be easily installed with pip:
 
 ```bash
-pip install -r requirements.txt
+pip install MultiMM
 ```
 
+PyPI software: https://pypi.org/project/MultiMM/.
+
 ## Input Data
 
 MultiMM relies on three types of datasets:
@@ -98,7 +100,7 @@ TRJ_FRAMES = 100
 After specifying the parameters and forces, users can run the following command in the terminal:
 
 ```bash
-python run.py -c config.ini
+MultiMM -c config.ini
 ```
 
 The software will output a folder with the resulting structure and plots showing compartment distribution.
@@ -107,6 +109,28 @@ Example data can be found on Google Drive: https://drive.google.com/drive/folder
 
 In the `examples` folder, we provide example configuration files for different modeling scenarios.
 
+## Vizualization
+
+For vizualization purposes, if you would like to import the whole genome structure you may use the command,
+
+```python
+import simulation.plots as splt
+
+splt.viz_chroms(sim_path)
+```
+
+For `sim_path` you should add the output folder directory path (add `comps=Fase` in case that you do not need compartment coloring). Otherwise, in case that the user would like to model a particular region, without using compartment or chromosome coloring (pretty much any cif structure), they can type,
+
+```python
+import simulation.plots as splt
+import simulation.utils as suts
+
+V = suts.get_coordinates_cif(cif_path)
+splt.viz_structure(V)
+``` 
+
+We would like once again to thank people who developed `pyvista` library and allow us to have fast and good vizualizations of large chromatin structure.
+
 ## Simulation Arguments
 
 MultiMM has numerous configurable parameters. Below is a description of each argument and its default values. The defaults have been tested for genome-wide simulation but can be modified in the configuration file if needed. Units are typically assumed based on OpenMM conventions, though explicit unit specification is not required.
@@ -121,7 +145,7 @@ This version should now be more precise and polished. Let me know if you'd like
 | DEVICE                       | str          | None        | None          | device index for CUDA or OpenCL (count from 0) |
 | INITIAL_STRUCTURE_PATH       | str          | None        | None          | Path to CIF file. |
 | BUILD_INITIAL_STRUCTURE      | bool         | True        | None          | To build a new initial structure. |
-| INITIAL_STRUCTURE_TYPE       | str          | hilbert     | None          | you can choose between: hilbert, circle. |
+| INITIAL_STRUCTURE_TYPE       | str          | hilbert     | None          | you can choose between: hilbert, circle, rw, confined_rw, self_avoiding_rw, helix, spiral, sphere, knot. |
 | FORCEFIELD_PATH              | str          | None        | None          | Path to XML file with forcefield. |
 | N_BEADS                      | int          | 50000       | None          | Number of Simulation Beads. |
 | COMPARTMENT_PATH             | str          | None        | None          | It should be  a `.bed` file with subcompartments from Calder. |
@@ -170,7 +194,6 @@ This version should now be more precise and polished. Let me know if you'd like
 | CF_USE_CENTRAL_FORCE         | bool         | False       | None          | Attraction of smaller chromosomes to the nucleolus (in the center). |
 | CF_STRENGTH                  | float        | 10.0       | kJ/mol        | Strength of Attraction of Smaller Chromosomes |
 | NUC_DO_INTERPOLATION         | bool         | False       | None          | Attraction of smaller chromosomes. |
-| MAX_NUCS_PER_BEAD            | int          | 4           | None          | Maximum amount of nucleosomes per single bead for nucleosome interpolation. |
 | NUC_RADIUS                   | float        | 0.1         | None          | The radius of the single nucleosome helix. |
 | POINTS_PER_NUC               | int          | 20          | None          | The number of points that consist a nucleosome helix. |
 | PHI_NORM                     | float        | pi/5        | None          | Zig zag angle. |

{multimm-1.0.5 → multimm-1.0.7}/setup.py

@@ -6,7 +6,7 @@ with open('README.md', 'r', encoding='utf-8') as f:
 
 setup(
     name='MultiMM',  # Package name
-    version='1.0.5',  # Version of the software
+    version='1.0.7',  # Version of the software
     description='A tool for chromatin modeling from nucleosomes to chromosomal territories.',
     long_description=long_description,
     long_description_content_type='text/markdown',

{multimm-1.0.5 → multimm-1.0.7}/simulation/__init__.py

@@ -1,4 +1,4 @@
 # Importing specific functions or classes from submodules
 from .run import main
 
-__version__ = "1.0.4"
+__version__ = "1.0.7"

{multimm-1.0.5 → multimm-1.0.7}/simulation/initial_structure_tools.py

@@ -136,33 +136,11 @@ def read_compartments(file,ch,reg,res,binary=True):
 def generate_hilbert_curve(n_points,p=8,n=3,displacement_sigma=0.1,scale=6,viz=False,add_noise=False):
     hilbert_curve = HilbertCurve(p, n)
 
-    distances = list(range(n_points)) # if n_points>4095 else list(range(n_points))
+    distances = list(range(n_points))
     points = np.array(hilbert_curve.points_from_distances(distances))
-    
-    #if n_points>4095:
-    #   x_sim, y_sim, z_sim = points[:,0], points[:,1], points[:,2]
-    #   tck, u = interpolate.splprep(x=[x_sim,y_sim,z_sim], s=2)
-    #   u_fine = np.linspace(0,1,n_points)
-    #   x_fine, y_fine, z_fine = interpolate.splev(u_fine, tck)
-    #   V_interpol = np.vstack((x_fine,y_fine,z_fine)).T
-    #else:
-    #   V_interpol = points
     if add_noise:
         displacement = np.random.normal(loc=0.0, scale=displacement_sigma, size=n_points*3).reshape(n_points,3)
         V_interpol = V_interpol + displacement
-    #max_dist = np.max(np.linalg.norm(V_interpol,axis=1))
-
-    #if viz:
-    #    fig = plt.figure()
-    #    ax = plt.axes(projection='3d')
-
-    #    z = V_interpol[:,2]
-    #    x = V_interpol[:,0]
-    #    y = V_interpol[:,1]
-
-    #    ax.plot3D (x, y, z, 'green')
-    #    ax.set_title('Hilbert Curve')
-    #    plt.show()
     
     return points
 
@@ -181,13 +159,97 @@ def polymer_circle(n: int, z_stretch: float = 1.0, radius: float = 5.0) -> np.nd
     points = np.array(points)
     return points
 
+def helix_structure(N_beads, radius=1, pitch=2):
+    theta = np.linspace(0, 4 * np.pi, N_beads)  # 2 full turns
+    x = radius * np.cos(theta)
+    y = radius * np.sin(theta)
+    z = np.linspace(0, pitch * N_beads, N_beads)
+    V = np.column_stack((x, y, z))
+    return V
+
+def spiral_structure(N_beads, initial_radius=1, pitch=1, growth_factor=0.05):
+    theta = np.linspace(0, 4 * np.pi, N_beads)
+    radius = initial_radius + growth_factor * np.arange(N_beads)
+    x = radius * np.cos(theta)
+    y = radius * np.sin(theta)
+    z = np.linspace(0, pitch * N_beads, N_beads)
+    V = np.column_stack((x, y, z))
+    return V
+
+def sphere_surface_structure(N_beads, radius=1):
+    phi = np.random.uniform(0, 2 * np.pi, N_beads)
+    costheta = np.random.uniform(-1, 1, N_beads)
+    u = np.random.uniform(0, 1, N_beads)
+    
+    theta = np.arccos(costheta)
+    r = radius * u ** (1/3)
+
+    x = r * np.sin(theta) * np.cos(phi)
+    y = r * np.sin(theta) * np.sin(phi)
+    z = r * np.cos(theta)
+    
+    V = np.column_stack((x, y, z))
+    return V
+
+def confined_random_walk(N_beads, box_size=5):
+    V = np.zeros((N_beads, 3))
+    for i in range(1, N_beads):
+        step = np.random.choice([-1, 1], size=3)  # Random step in x, y, z
+        V[i] = V[i-1] + step
+        # Keep the points within a confined box (folding in)
+        V[i] = np.clip(V[i], -box_size, box_size)
+    return V
+
+def trefoil_knot_structure(N_beads, scale=5):
+    t = np.linspace(0, 2 * np.pi, N_beads)
+    x = scale * (np.sin(t) + 2 * np.sin(2 * t))
+    y = scale * (np.cos(t) - 2 * np.cos(2 * t))
+    z = -scale * np.sin(3 * t)
+    
+    V = np.column_stack((x, y, z))
+    return V
+
+def random_walk_structure(N_beads, step_size=1):
+    # Initialize the structure array
+    V = np.zeros((N_beads, 3))
+    
+    # Loop over each bead, starting from the second one
+    for i in range(1, N_beads):
+        # Randomly pick a direction for each step
+        step_direction = np.random.normal(size=3)
+        step_direction /= np.linalg.norm(step_direction)  # Normalize to make it unit length
+        
+        # Move the current bead from the last bead by a fixed step size
+        V[i] = V[i-1] + step_size * step_direction
+    
+    return V
+
+def compute_init_struct(N_beads,mode='rw',scale=5):
+    match mode:
+        case 'rw':
+            return random_walk_structure(N_beads)
+        case 'confined_rw':
+            return confined_random_walk(N_beads)
+        case 'knot':
+            return trefoil_knot_structure(N_beads)
+        case 'self_avoiding_rw':
+            return self_avoiding_random_walk(N_beads)
+        case 'circle':
+            return polymer_circle(N_beads,50,5)
+        case 'helix':
+            return helix_structure(N_beads)
+        case 'spiral':
+            return spiral_structure(N_beads)
+        case 'sphere':
+            return sphere_surface_structure(N_beads)
+        case 'hilbert':
+            return generate_hilbert_curve(n_dna,scale=scale)
+        case _:
+            return IndentationError('Invalid option for initial structure.')
+
 def build_init_mmcif(n_dna,chrom_ends,psf=True,path='',curve='hilbert',scale=5):
     # Define the initial coordinates of histones and the structure of DNA
-    if curve=='hilbert':
-        dna_points = generate_hilbert_curve(n_dna,scale=scale) 
-    elif curve=='circle': 
-        dna_points = polymer_circle(n_dna,50,5)
-    
+    dna_points = compute_init_struct(n_dna,mode=curve,scale=scale)
     # Write the positions in .mmcif file
     atoms = ''
     

{multimm-1.0.5 → multimm-1.0.7}/simulation/nucleosome_interpolation.py

@@ -29,9 +29,9 @@ def get_perp_component(x, v):
 
 
 class NucleosomeInterpolation:
-    def __init__(self, V, bw, max_nucs_per_bead=4, nuc_radius=0.1, points_per_nuc=20, phi_norm=np.pi/5):
+    def __init__(self, V, bw, nuc_radius=0.1, points_per_nuc=20, phi_norm=np.pi/5):
         self.V, self.bw = V, bw
-        self.max_nucs_per_bead, self.nuc_points = max_nucs_per_bead, points_per_nuc
+        self.max_nucs_per_bead, self.nuc_points = int(np.ceil(1/(2*nuc_radius))), points_per_nuc
         self.nuc_r, self.phi_norm = nuc_radius, phi_norm
     
     def make_helix(self, r, theta, z0):