Klauda Lab Simulation Wiki

User Tools

Site Tools

Trace:
Writing /users/jbklauda/pub/wiki/data/cache/6/6193a558dc90ed4c190004bf67ecff70.captchaip failed
packing_defects

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
Next revision		Previous revision
packing_defects [2017/05/21 22:46] – [Step 1: VMD Defect imaging and coordinate data collection] editpacking_defects [2019/11/18 11:54] (current) – admin
Line 2: Line 2:
  
 ==== Packing Defect Analysis Overview ==== ==== Packing Defect Analysis Overview ====
 +
 +**Please Cite Source**: Wildermuth, K.; Monje-Galvan, V.; Warburton, L. M.; Klauda, J. B., Effect of membrane lipid packing on stable binding of the ALPS peptide. //Journal of Chemical Theory and Computation// 2019, **15**, 1418-1429. ([[https://pubs.acs.org/doi/10.1021/acs.jctc.8b00945]])
  
 This analysis focuses on the study of membrane lipid packing defects -- areas of low density lipid packing on the surface of a membrane that expose the nonpolar core of the membrane. Packing defects arise from dispersion of membrane constituents due to external forces such as imposed surface tensions or chemical interactions with other molecules. The analysis of packing defects is useful when quantifying the membrane surface is of interest. For example, I used packing defect quantification to describe the dynamic binding event between the Osh4 ALPS-like peptide and a membrane surface, and to describe the effects of using the HMMM simulation technique.  This analysis focuses on the study of membrane lipid packing defects -- areas of low density lipid packing on the surface of a membrane that expose the nonpolar core of the membrane. Packing defects arise from dispersion of membrane constituents due to external forces such as imposed surface tensions or chemical interactions with other molecules. The analysis of packing defects is useful when quantifying the membrane surface is of interest. For example, I used packing defect quantification to describe the dynamic binding event between the Osh4 ALPS-like peptide and a membrane surface, and to describe the effects of using the HMMM simulation technique. 
Line 24: Line 26:
   - **DCD trajectory file:** 1 DCD file containing the trajectory to be analyzed. It is recommended to use a DCD with a frame count corresponding to the number of images desired to be outputted (use [[merge DCD]]), although there are parameters available in this script for skipping frames in the DCD.    - **DCD trajectory file:** 1 DCD file containing the trajectory to be analyzed. It is recommended to use a DCD with a frame count corresponding to the number of images desired to be outputted (use [[merge DCD]]), although there are parameters available in this script for skipping frames in the DCD. 
   - **PSF file:** associated PSF file for the trajectory.    - **PSF file:** associated PSF file for the trajectory. 
-  - **[[Defect_tachyon_scav.tcl]]:** script called on by VMD +  - **[[defect_tachyon_scav.tcl]]:** script called on by VMD 
-  - **[[Defect_tachyon.csh]]:** c-shell script to submit job with SLURM and call VMD (default setup for DT1 operation)+  - **[[defect_tachyon.csh]]:** c-shell script to submit job with SLURM and call VMD (default setup for DT1 operation)
  
 ===Output files=== ===Output files===
   - **Image files:** Image files (default BMP) are saved within ''home'' using an automated directory architecture created by the script. Two image sets (protein and lipid) for each leaflet of the membrane are created, separated into two directories within ''home'' (top, bottom) for a total of four image sets. An optional side directory will be created if this image set is selected to be written.    - **Image files:** Image files (default BMP) are saved within ''home'' using an automated directory architecture created by the script. Two image sets (protein and lipid) for each leaflet of the membrane are created, separated into two directories within ''home'' (top, bottom) for a total of four image sets. An optional side directory will be created if this image set is selected to be written. 
-  - **Data files:** text files are also created that include coordinate adta to be used in associated analysis scripts. +  - **Data files:** text files are also created that include coordinate data to be used in associated analysis scripts. 
     - **pbc_scale_dimnesions.dat:** PBC coordinates of first frame of DCD     - **pbc_scale_dimnesions.dat:** PBC coordinates of first frame of DCD
     - **PBC_coord.dat:** contain PBC coordinates     - **PBC_coord.dat:** contain PBC coordinates
Line 35: Line 37:
     - **prot_coord.dat, prot_minmax_coord.dat:** contain time series associated with protein     - **prot_coord.dat, prot_minmax_coord.dat:** contain time series associated with protein
     - **custom_coord.dat:** time series for a defined custom selection (typically a protein residue of interest)     - **custom_coord.dat:** time series for a defined custom selection (typically a protein residue of interest)
-    - **VMD output file:** will be written directly to ''home''. Contains output lines from VMD operation. Useful for troublshooting the script. +    - **VMD output file:** will be written directly to ''home''. Contains output lines from VMD operation. Useful for troubleshooting the script. 
  
 ===Set up instructions=== ===Set up instructions===
Line 42: Line 44:
     - The polar portion of some lipids have been predefined at the top of the script (ex. ''polar_PS'', ''polar_PE''). If a lipid in your system is not already listed, it must be defined. Default definition of polar is to include all atoms in the head group down to and including the second carbon in the lipid chain.      - The polar portion of some lipids have been predefined at the top of the script (ex. ''polar_PS'', ''polar_PE''). If a lipid in your system is not already listed, it must be defined. Default definition of polar is to include all atoms in the head group down to and including the second carbon in the lipid chain. 
     - If using an HMMM membrane, a selection exists to label all DCLE atoms nonpolar that are contained within the core of the membrane (default is -20Å<z<+20Å since some DCLE molecules escape the membrane)     - If using an HMMM membrane, a selection exists to label all DCLE atoms nonpolar that are contained within the core of the membrane (default is -20Å<z<+20Å since some DCLE molecules escape the membrane)
-    - The default script draws the representations using the Quicksurf option. Quicksurf is relatively computationally expensive. For example, in my work I used quicksurf and image creation took a few days. I tried VDW and it took less than an hour. I think using VDW representation will be OK (this needs to be confirmed) but **make sure to use the same representation for all analysis to maintain consistency during analysis!**+    - The default script draws the representations using the Quicksurf option. Quicksurf is relatively computationally expensive. For example, in my work I used quicksurf and image creation took a few days. I tried VDW spheres and it took less than an hour. I think using VDW representation will be OK (this needs to be confirmed) but **make sure to use the same representation for all analysis to maintain consistency during analysis!**
   - Modify the ''custom_sel'' variable if you want to track a specific selection during the trajectory. For example, if you wanted to track the TRP alpha carbon of a peptide, you could set custom_sel atom selection to ''resname TRP and name CA''. The associated analysis scripts are constructed to read a custom_sel file, thus if you do not want to track a custom selection, just in the selection with something arbitrary and ignore its output.    - Modify the ''custom_sel'' variable if you want to track a specific selection during the trajectory. For example, if you wanted to track the TRP alpha carbon of a peptide, you could set custom_sel atom selection to ''resname TRP and name CA''. The associated analysis scripts are constructed to read a custom_sel file, thus if you do not want to track a custom selection, just in the selection with something arbitrary and ignore its output. 
   - set ''xsize'', ''ysize'': pixel dimensions of outputted images (default 600x600), recommended to keep 1:1 ratio.   - set ''xsize'', ''ysize'': pixel dimensions of outputted images (default 600x600), recommended to keep 1:1 ratio.
   - set ''home'': filepath to directory that contains input files   - set ''home'': filepath to directory that contains input files
-  - This script uses PBCTools (http://www.ks.uiuc.edu/Research/vmd/plugins/pbctools/) to draw a box around the boundary of the system, centered on the protein. +  - This script uses [[http://www.ks.uiuc.edu/Research/vmd/plugins/pbctools/|PBCTools]] to draw a box around the boundary of the system, centered on the protein. 
     - If no protein in system, can keep box centered on origin of system. The definition of the box needs to be edited if running the code on a system without a protein.      - If no protein in system, can keep box centered on origin of system. The definition of the box needs to be edited if running the code on a system without a protein. 
     - Troubleshooting: f the protein wraps when viewed in VMD, uncomment the following line in the script to rewrap the membrane centered on the protein      - Troubleshooting: f the protein wraps when viewed in VMD, uncomment the following line in the script to rewrap the membrane centered on the protein 
Line 61: Line 63:
  
 ===Running instructions=== ===Running instructions===
-From command line in ''home'', call ''defect_tachyon.csh''. The csh file calls vmd. +From command line in ''home'', call ''defect_tachyon.csh''. The cshell file calls VMD. 
   * The default file contains the filepath to the executable link for VMD on DT1, this should be changed if using a different system, or a newer version of VMD.   * The default file contains the filepath to the executable link for VMD on DT1, this should be changed if using a different system, or a newer version of VMD.
-  * The csh file calls vmd without a display, so this process will run in the background +  * The chsell script calls VMD without a display, so this process will run in the background 
-  * The default file runs this process using the scavenger partition. You can change to another partition but this process runs fine on scavenger. I designed the code to pick up from where the process left off if it is cancelled in the middle of the run, so if you are knocked off, simply recall the csh file and the job will continue from where it left off. Though I have never been knocked off scavenger. +  * The default file runs this process using the scavenger partition. You can change to another partition but this process runs fine on scavenger. I designed the code to pick up from where the process left off if it is cancelled in the middle of the run, so if you are knocked off, simply recall the cshell file and the job will continue from where it left off. Though I have never been knocked off scavenger. 
  
  
Line 71: Line 73:
  
 === Purpose === === Purpose ===
-Now we will run an image processing script on the prepared image sets. This step utilizes python and a really neat computer vision library called OpenCV http://opencv.org/). Both softwares are available on DT1, and these directions will assume use of DT1 for operation of the analysis. Though in theory, this analysis can be run on any system with Python (and numpy) and OpenCV. Furthermore, I assume the use of OpenCV v.3.1.0 and Python v.2.7.8+Now we will run an image processing script on the prepared image sets. This step utilizes python and a computer vision library called [[http://opencv.org/|OpenCV]]. Both softwares are available on DT1, and these directions will assume use of DT1 for operation of the analysis. Though in theory, this analysis can be run on any system with Python (and numpy) and OpenCV. Furthermore, I assume the use of OpenCV v.3.1.0 and Python v.2.7.8
  
 A visual check of some of the images should be conducted before running the code to make sure things are in order. For example, make sure the green box is visible in the frame and, if there is a protein in the system, make sure the green box is centered on the protein. The protein should not wrap around the edge of the PBC. Protein images should contain only the protein, and lipid images should contain only the lipid.  A visual check of some of the images should be conducted before running the code to make sure things are in order. For example, make sure the green box is visible in the frame and, if there is a protein in the system, make sure the green box is centered on the protein. The protein should not wrap around the edge of the PBC. Protein images should contain only the protein, and lipid images should contain only the lipid. 
  
 === Input files === === Input files ===
-  - **blob_top.py, blob_bot.py:** These are the python script files. Both files should be in ''home'' which contains the directories built by the previous vmd script. Note that the files are very similar, though ''blob_top.py'' runs analysis for the top leaflet of the membrane and ''blot_bot.py'' runs analysis for the bottom leaflet of the membrane +  - **[[blob_top.py]], [[blob_bot.py]]:** These are the python script files. Both files should be in ''home'' which contains the directories built by the previous VMD script. Note that the files are very similar, though ''blob_top.py'' runs analysis for the top leaflet of the membrane and ''blot_bot.py'' runs analysis for the bottom leaflet of the membrane 
   - **pbc_scale_dimensions.dat and image sets**: these are built by the previous VMD script and are automatically structured appropriately.    - **pbc_scale_dimensions.dat and image sets**: these are built by the previous VMD script and are automatically structured appropriately. 
  
 === Output files === === Output files ===
-  - **local_defect_top.dat, local_defect_bot.dat:** These files contain information about the packing defects closest to binding protein (if the file is edited to analyze membranes without proteins, this output file should be removed). The output files contain columns for frame number of the trajectory, area of defect below the protein (in Å^2), area of protein viewed from normal of z-plane (in Å^2), and a protein comparison value which is smaller when the local packing defect shape is similar to the protein shape (see openCV matchShapes function for more information http://docs.opencv.org/2.4/modules/imgproc/doc/structural_analysis_and_shape_descriptors.html?highlight=drawcontours#double matchShapes(InputArray contour1, InputArray contour2, int method, double parameter)) +  - **local_defect_top.dat, local_defect_bot.dat:** These files contain information about the packing defects closest to binding protein (if the file is edited to analyze membranes without proteins, this output file should be removed). The output files contain columns for frame number of the trajectory, area of defect below the protein (in Å^2), area of protein viewed from normal of z-plane (in Å^2), and a protein comparison value which is smaller when the local packing defect shape is similar to the protein shape (see openCV [[http://docs.opencv.org/2.4/modules/imgproc/doc/structural_analysis_and_shape_descriptors.html?highlight=drawcontours#double matchShapes(InputArray contour1, InputArray contour2, int method, double parameter))|matchShapes]] function for more information  
-  - **global_defect_top.dat, global_defect_bot.dat:** These files contain information about the overall packing defect status of the entire system. The output fules contain columns for frame number of the trajectory, area of the entire leaflet (in Å^2), and area of total defect on the leaflet (in Å^2). +  - **global_defect_top.dat, global_defect_bot.dat:** These files contain information about the overall packing defect status of the entire system. The output files contain columns for frame number of the trajectory, area of the entire leaflet (in Å^2), and area of total defect on the leaflet (in Å^2). 
   - **defect_freq_area_top.dat, defect_freq_area_bot.dat:** These files contain a list of all identified packing defects on their respective membrane leaflets. The columns from left to right are frame number that the defect appears in, the area of the defect (in Å^2), and the protein comparison value for that defect relative to the protein during the same frame.   - **defect_freq_area_top.dat, defect_freq_area_bot.dat:** These files contain a list of all identified packing defects on their respective membrane leaflets. The columns from left to right are frame number that the defect appears in, the area of the defect (in Å^2), and the protein comparison value for that defect relative to the protein during the same frame.
     - New defect properties can be defined in the script and added to this output file        - New defect properties can be defined in the script and added to this output file   
Line 117: Line 119:
  
  
-=== Protein Binding: Conditional Porbability Analysis ===+=== Protein Binding: Conditional Probability Analysis ===
  
 Some proteins bind to membranes at large packing defects. One way to show that this is true is by comparing the probability of the peptide moving towards the membrane to the conditional probability of the peptide moving towards the membrane given that a large packing defect exists near the peptide (theoretical details are outlined in a paper published by our group). Using information from the packing defect analysis, we can run a script to compute the necessary probabilities to conduct this analysis.  Some proteins bind to membranes at large packing defects. One way to show that this is true is by comparing the probability of the peptide moving towards the membrane to the conditional probability of the peptide moving towards the membrane given that a large packing defect exists near the peptide (theoretical details are outlined in a paper published by our group). Using information from the packing defect analysis, we can run a script to compute the necessary probabilities to conduct this analysis. 
  
-The script ''[[truDiff.py]]'' can be run in the ''data_files'' directory created by the VMD script. This script will measure the minimum z-distance between the protein and the phosphate layer of a specified leaflet of the membrane. In the script, specify the final binding ''leaflet'' ('top' or 'bottom'), and the number of frames in the dcd (''num_frame'' and ''prebind_limit''). Execute the script from the command line by using the following code while in the ''data_files'' directory:+The script ''[[truDiff.py]]'' can be run in the ''data_files'' directory created by the VMD script. This script will measure the minimum z-distance between the protein and the phosphate layer of a specified leaflet of the membrane. In the script, specify the final binding ''leaflet'' ('top' or 'bottom'), and the number of frames in the DCD (''num_frame'' and ''prebind_limit''). Execute the script from the command line by using the following code while in the ''data_files'' directory:
 <code> <code>
 module load python/2.7.8 module load python/2.7.8
 python truDiff.py python truDiff.py
 </code>  </code> 
-There will be a short prompt and the computation will compelte quickly. An output file ''binding_distances.dat'' will be made which contains the necessary minimum binding distance time series, as well two other distance time series described in the script's header+There will be a short prompt and the computation will complete quickly. An output file ''binding_distances.dat'' will be made which contains the necessary minimum binding distance time series, as well two other distance time series described in the script's header
  
-Next, the script ''[[defect_peptide_frequency.py]]'' will count the number of times the peptide moves towards the membrane by a specified distance (m), how many instances a defect exists larger than a specified area (d), and how many times both of those instances occur simoltaneously. These frequencies can be used to compute the probabilities and coniditional probabilities described above. Options in the script allow the user to set the range of defect area and movement variable space, distance cutoffs (so you can analyze interactions only in a specific distance range between the membrane and the peptide), and trajectory cutoffs (so you can analyze interactions only in a specific time range, for example while the peptide is unbinded). Choose ''freq_path'' to be the ''local_defect'' data file for the leaflet of interest (typically the binding leaflet) and make sure the ''dist_path'' is set to the truDiff output file. Execute the script from the command line by using the following code while in the ''data_files'' directory:  +Next, the script ''[[defect_peptide_frequency.py]]'' will count the number of times the peptide moves towards the membrane by a specified distance (m), how many instances a defect exists larger than a specified area (d), and how many times both of those instances occur simultaneously. These frequencies can be used to compute the probabilities and conditional probabilities described above. Options in the script allow the user to set the range of defect area and movement variable space, distance cutoffs (so you can analyze interactions only in a specific distance range between the membrane and the peptide), and trajectory cutoffs (so you can analyze interactions only in a specific time range, for example while the peptide is unbinded). Choose ''freq_path'' to be the ''local_defect'' data file for the leaflet of interest (typically the binding leaflet) and make sure the ''dist_path'' is set to the truDiff output file. Execute the script from the command line by using the following code while in the ''data_files'' directory:  
 <code>  <code> 
 module load python/2.7.8 module load python/2.7.8
 python defect_peptide_frequency.py python defect_peptide_frequency.py
 </code> </code>
-  * Note, the ''module load '' comand does not have to be executed if the python module has already been loaded. +  * Note, the ''module load '' command does not have to be executed if the python module has already been loaded. 
   * The script will run in the command line, and takes a few minutes. The speed of the script is dependent primarily on the size of the movement variable sample space.    * The script will run in the command line, and takes a few minutes. The speed of the script is dependent primarily on the size of the movement variable sample space. 
   * The output file ''freq_stat_total.dat'' will contain counts of each event described above in a large table. Probabilities can be calculated from these tallies, and various results plotted (insert example excel sheet)   * The output file ''freq_stat_total.dat'' will contain counts of each event described above in a large table. Probabilities can be calculated from these tallies, and various results plotted (insert example excel sheet)
packing_defects.1495421208.txt.gz · Last modified: 2017/05/21 22:46 by edit