Merge branch 'main' of https://github.com/kiharalab/DiffModeler

Author: Xiao Wang

README.md

@@ -20,19 +20,21 @@ For technical problems or questions, please reach to Xiao Wang (wang3702@purdue.
 
 ## Citation:
 
-Xiao Wang, Han Zhu, Genki Terashi & Daisuke Kihara. Protein Complex Structure Modeling with Diffusion Model and AlphaFold in cryo-EM maps.bioArxiv, 2023.
+Xiao Wang, Han Zhu, Genki Terashi & Daisuke Kihara. DiffModeler: Large Macromolecular Structure Modeling for Cryo-EM Maps Using Diffusion Model. Nature Methods, accepted 2024.<br>
+Early Version available at [bioRxiv](https://www.biorxiv.org/content/10.1101/2024.01.20.576370v2)
 ```
 @article{wang2023DiffModeler,   
-  title={Protein Complex Structure Modeling with Diffusion Model and AlphaFold in cryo-EM maps},   
-  author={Xiao Wang, Han Zhu, Genki Terashi, and Daisuke Kihara},    
-  journal={bioArxiv},    
-  year={2023}    
+  title={DiffModeler: Large Macromolecular Structure Modeling for Cryo-EM Maps Using Diffusion Model},   
+  author={Xiao Wang, Han Zhu, Genki Terashi, Manav Taluja, and Daisuke Kihara},    
+  journal={Nature Methods},    
+  year={2024}    
 }   
 ```
 
 ## Free Online Server: 
 ### Input map+single-chain structures: https://em.kiharalab.org/algorithm/DiffModeler
 ### Input map+sequence: https://em.kiharalab.org/algorithm/DiffModeler(seq)
+### Protein+DNA Complex structure modeling: https://em.kiharalab.org/algorithm/ComplexModeler
 
 ## Introduction
 

data_processing/Unify_Map.py

@@ -13,6 +13,7 @@ def Unify_Map(input_map_path, new_map_path):
     nstart = np.asarray([mrc.header.nxstart, mrc.header.nystart, mrc.header.nzstart], dtype=np.float32)
     cella = np.array(mrc.header.cella.tolist(), dtype=np.float32)
     mapcrs = np.asarray([mrc.header.mapc, mrc.header.mapr, mrc.header.maps], dtype=int)
+    msample = np.asarray([mrc.header.mx, mrc.header.my, mrc.header.mz], dtype=int)
     # if np.sum(nstart) == 0:
     #     return input_map_path
     mrc.print_header()
@@ -26,13 +27,14 @@ def Unify_Map(input_map_path, new_map_path):
     data = np.transpose(data, axes=2 - sort[::-1])
 
     # Move offsets from nstart to origin if origin is zero (MRC2000)
-    if np.sum(origin) == 0:
+    if np.all(origin == 0):
         origin = origin + nstart * voxel_size
 
     # Save the unified map
     mrc_new = mrcfile.new(new_map_path, data=data, overwrite=True)
     (mrc_new.header.origin.x, mrc_new.header.origin.y, mrc_new.header.origin.z) = origin
     (mrc_new.header.cella.x, mrc_new.header.cella.y, mrc_new.header.cella.z) = cella
+    (mrc_new.header.mx, mrc_new.header.my, mrc_new.header.mz) = msample
     mrc_new.update_header_stats()
     mrc_new.print_header()
     mrc_new.close()

ops/pdb2vol.py

@@ -183,10 +183,10 @@ def mapGridPosition(origin, voxel_size, box_size, atom_coord):
     y_pos = (atom_coord[1] - origin[1]) / voxel_size[1]
     z_pos = (atom_coord[2] - origin[2]) / voxel_size[2]
 
-    if (box_size[2] > x_pos >= 0) and (box_size[1] > y_pos >= 0) and (box_size[0] > z_pos >= 0):
+    if (box_size[2] > x_pos + 1 >= 0) and (box_size[1] > y_pos + 1 >= 0) and (box_size[0] > z_pos + 1 >= 0):
         return x_pos, y_pos, z_pos
     else:
-        return 0, 0, 0
+        return 0.0, 0.0, 0.0
 
 
 @njit(fastmath=True, nogil=True)
@@ -207,7 +207,7 @@ def make_atom_overlay_map(origin, voxel_size, box_size, atom_list, atom_type_lis
     map_data = np.zeros(box_size)
     for atom, atom_type in zip(atom_list, atom_type_list):
         pos = mapGridPosition(origin, voxel_size, box_size, atom)
-        if pos:
+        if not (pos[0] == pos[1] == pos[2] == 0.0):
             atom_mass = atom_mass_dict.get(atom_type, 0.0)
             pos_x_0 = int(np.floor(pos[0]))
             pos_y_0 = int(np.floor(pos[1]))