NYSBC course: nextPYP practical (day 1)¶

• Click on Create new project, give the project a name, and select Create

• Select the new project from the Dashboard and click Open

• The newly created project will be empty and a Jobs panel will appear on the right

• Go to Import Data and select Tomography (from Raw Data)

• A form to enter parameters will appear:

• Go to the Raw data tab:

  • Set the path to raw data by clicking on the icon and browse to /nfs/bartesaghilab/nextpyp/workshop/10164/

  • Type *.tif into the filter box (lower right) and click the icon

• Go the the Microscope Parameters tab:

  • Set Pixel size (A) to 1.35

  • Set Acceleration voltage (kV) to 300

  • Set Tilt-axis angle (degrees) to 85.3

• Click Save and the new block will appear on the project page. The block is in the modified state (indicated by the sign) and is ready to be executed

• Clicking the Run button will show another dialog where you can select which blocks to run:

• Click Start Run for 1 block. This will launch a process that reads one tilt at random and displays the resulting image inside the block

• Click on the thumbnail inside the block to see a larger version of the projection image

• Click on Tilt-series (output of the Tomography (from Raw Data) block) and select Pre-processing

• Go to the Frame alignment tab:

  • nextPYP uses the Frame pattern to extract metadata form the file names. EMPIAR-10164 follows the default file naming scheme and .tif extension, so we will leave the default setting.

  • We will use unblur for frame alignment.

• Go to the CTF determination tab

  • Set Max resolution to 5

• Go to the Tilt-series alignment tab

  • Our Alignment method will be IMOD fiducial-based which is the default so make no changes.

• Go to the Tomogram reconstruction tab

  • Our Reconstruction method will be IMOD, this is the default so make no changes.

• Go to the Resources tab

  • Set Split, Threads to 41

• Click Save, Run, and Start Run for 1 block. Follow the status of the run in the Jobs panel

• When the block finishes running, examine the Tilt-series, Plots, Table, and Gallery tabs. We will measure our virions in this block as well.

• We will be utilizing three separate blocks to perform geometrically constrained particle picking. This will allow for increased accruacy in particle detection and provides geometric priors for downstream refinement.

• Block 1: Virion selection

  • Click on Tomograms (output of the Pre-processing block) and select Particle-Picking

  • Go to the Particle detection tab:

    • Set Detection method to virions

    • Set Virion radius (A) to 500 (half the diameter we measured)

  • Click Save

• Block 2: Virion segmentation

  • Click on Particles (output of the Particle-Picking block) and select Segmentation (closed surfaces)

  • Click Save

• Block 3: Spike (Gag) detection

  • Click on Segmentation (closed) (output of the Segmentation (closed surfaces) block) and select Particle-Picking (closed surfaces)

  • Go to the Particle detection tab:

    • Set Detection method to uniform

    • Set Particle radius (A) to 50

    • Set Size of equatorial band to restrict spike picking (A) to 800

  • Click Save, Run, and Start Run for 3 blocks. Follow the status of the run in the Jobs panel

• In the upper left of your project page, click Import Workflow

• Choose the 2025 NYSBC course: Pre-processing (EMPIAR-10499) workflow by clicking the Import button to its right

• We pre-set the parameters for the workflow, so you can immediately click Save. Three blocks will populate on the project page.

• Click into the settings of the Particle-Picking block

  • Set Particle radius (A) to 80

  • Change Detection method from none to size-based using the dropdown menu

• Click Save, Run, and Start Run for 3 blocks. Follow the status of the run in the Jobs panel

• Click on the menu for the Particle-Picking block

• Select Copy

• Check Copy files and data and Make automatically-picked particles editable

• Click Next

• Click into the new Particle-Picking block.

• Ensure you are on the Particles tab. Here, you can right click to remove particles and left click to add particles.

• This manual picking feature is what I used the generate a particle set for nn-training for the next particle picking method we will use on the third dataset.

• In the upper left of your project page, click Import Workflow

• Choose the 2025 NYSBC course: Pre-processing (EMPIAR-10987) workflow by clicking Import

• We pre-set the parameters for the workflow, so you can immediately click Save. Three blocks will populate on the project page.

• Click into the settings of the Particle-Picking (eval) block

  • Click the icon. Browse to /nfs/bartesaghilab/nextpyp/workshop/10987/model_last_contrastive.pth

  • Set Particle radius (A) to 100

  • Set Threshold for soft/hard positives to 0.5

  • Set Max number of particles to 700

• Click Save, Run, and Start Run for 3 blocks. Follow the status of the run in the Jobs panel

• Click on Tomograms (output of the Pre-processing block) and select Particle-Picking

• Set Detection method to import

• Set Particle radius (A) to 50

• Click and browse to /nfs/bartesaghilab/nextpyp/workshop/10164/particles. Select Choose Folder

• Click Save, Run, and Start Run for 1 block

• Click on Particles (output of the Particle-Pickng block) and select Calculate reconstruction

• Go to the Sample tab

  • Set Molecular weight (kDa) to 300

  • Set Particle radius (A) to 150

  • Set Symmetry to C6

• Go to the Extraction tab

  • Set Box size (pixels/voxels) to 128

  • Set Image binning to 2

• Go to the Alignments tab

  • From the Import from dropdown menu, select nextPYP (*.bz2)

  • Click the icon next to Input parameter file (*.bz2) and browse to /nfs/bartesaghilab/nextpyp/workshop/10164/tomo-coarse-refinement-fg2v2MJLSY4Ui908_r01_02.bz2 Click Choose File

• Go to the Reconstruction tab

  • Select Apply dose weighting by checking the box

• Go to the Resources tab

  I’m a core course participant I’m an additional TA

  • Set Split, Threads to 124

  • Set Split, Threads to 70

• Click Save, Run, and Start Run for 1 block

  

• Click on Particles (output of the Particle refinement block) and select Particle filtering

• Go to the Particle filtering tab

  • Set Score threshold to 3.5

  • Set Min distance between particles (unbinned pixels) to 54

  • Click the icon next to Input parameter file(*.bz2) and select the *.bz2 file that appears (this is from the parent directory). Click Choose File

  • Check the box next to Permanently remove particles

• Click Save, Run, and Start Run for 1 block

• Click on Particles (output of the Particle filtering block) and select 3D refinement

• Go to the Extraction tab

  • Set Box size (pixels/voxels) to 256

  • Set Image binning to 1

• Go to the Particle scoring function tab

  • Set Last tilt for refinement to 8

  • Set Max resolution (A) to 4:3.5

  • From the Masking strategy dropdown menu, select from file

  • Click the icon to select the Shape mask (*.mrc), browse to /nfs/bartesaghilab/nextpyp/workshop/10164/EMPIAR-10164_shape_mask.mrc, and click Choose File

• Go to the Refinement tab

  • Next to Input parameter file (*.bz2) click the icon. Select the _r01_02_clean.bz2 file and click Choose File

  • Set Last iteration to 3

  • Check Refine tilt-geometry

  • Check Refine particle alignments

  • Set Number of regions to 8,8,2

• Go to the Reconstruction tab

  • Check Apply dose weighting (It may already be checked)

• Click Save, Run, and Start Run for 1 block

  

• Click on Particles (output of the Particle refinement block) and select Movie refinement

• Go to the Particle scoring function tab

  • Set Last exposure for refinement to 4

  • Set Max resolution (A) to 3.5

• Go to the Frame refinement tab

  • Next to Input parameter file (*.bz2) click the icon. Select the _r01_03.bz2 file and click Choose File

  • Set Spatial sigma to 400

  • Set Time sigma to 16

• Go to the Reconstruction tab

  • Check Apply dose weighting

• Click Save, Run, and Start Run for 1 block

  

• For reference, these instructions are also available on the User Guide.

• We assume the user already has the ArtiaX plugin, if not a simple google search will bring you to their docs for installation.

• Download files

  • Select a tomogram you wish to visualize the particles in. I will be using TS_43.

  • Click into the Pre-processing block, go to Tilt Series tab and Tomogram sub tab. On this page, click the search icon, search for TS_43. Click the green button immediately above the tomogram display. This will download the tomogram in .rec format.

  • Click into the Particle refinement block, go to the Metadata tab. On this page, type TS_43 into the search bar and click Search. Click the .star file to download particle alignments.

  • Go to the Reconstruction tab and download the Cropped Map.

• Display in ChimeraX

  • Open ChimeraX (again, we assume ArtiaX is installed)

  • Open the tomogram TS_43.rec

  • Run the following commands in the ChimeraX shell:

  volume permuteAxes #1 xzy
  volume flip #2 axis z
  

  • Go to the ArtiaX tab and click Launch to start the plugin.

  • In the Tomograms section on the left, select model #3 (permuted z flip) from the Add Model dropdown menu and click Add!

  • Go to the ArtiaX options panel on the right, and set the Pixel Size for the Current Tomogram to 10.8 (The current binned pixel size)

  • On the left panel, under the Particles List section, select Open List … and open the .star file.

  • Return to the panel on the right and select the Select/Manipulate tab. Set the Origin to 1.35 (the unbinned pixel size)

  • From the Color Settings section, select Colormap and then rlnLogLikelihoodContribution from the dropdown menu.

  • Play with the Marker Radius and Axes Size sliders to visualize the particle locations, cross correlation scores, and orientations.

• Click on Frames (output of the Movie refinement block) and select Post-processing

• Go to the Post-processing tab

  • Next to First half map (*_half1.mrc) click the icon. Select the *_half1.mrc file and click Choose File

  • Set Masking method to from file usign the dropdown menu

  • Next to Mask file (*.mrc) click the icon. Browse to /nfs/bartesaghilab/nextpyp/workshop/10164/EMPIAR-10164_shape_mask.mrc and click Choose File

  • Set the B-factor method to adhoc using the dropdown menu

  • Set the Adhoc value (A^2) to -25

• Click Save, Run, and Start Run for 1 block

• I will be using a prealigned pdb file and files downloaded from nextPYP to demonstrate how one can visualize their final map aligned to a model in Chimera.

• Download files

  • In the Post-processing block, go to the Reconstruction tab. Click on the drop down menu Select an MRC file to download. Select the Full-Size Map. Your browser will download the post processed map as an MRC file.

  • We are using a pre-aligned, pre-cropped pdb file (5L93) so do not need to download this. For your experiments, you would download whatever model required.

  • Open the downloaded MRC file in Chimera. Visualize your beautiful map. To get a better look at your map/model fitting, open an atomic model in Chimera. Under the Map tab, Click Zone. Note we are left with a slightly larger zone than we would like so we will copy the zone command from the output to the terminal line, and edit the range. This leaves us with:

  volume zone #2 nearAtoms #1 range 2.4
  

  • Select the model, go to Actions, Atoms/Bonds, and Show Sidechain/Base

  • You can now view the model fit to your map interactively in ChimeraX

In this session we learned some of the things we are capable of doing in nextPYP:

• Raw data import

• Pre-processing (frame alignment, tilt-series alignment, CTF estimation)

• Tomogram reconstruction (WBP, fakeSIRT, SART)

  • nextPYP also supports tomogram denoising using cryoCARE, IsoNet and Topaz Denoise

• Segmentation (closed surfaces)

  • nextPYP also supports open surface segmentation which uses membrain-seg

• Particle picking (geometrically constrained, size-based, nn-based, manual)

  • nextPYP also supports template-search and molecular pattern mining

• Particle refinement (constrained single particle tomography, particle filtering, exposure weighting, region-based refinement, movie frame refinement, and post-processing)

  • nextPYP also supports particle-based CTF refinement, building shape masks, ab-initio refinement, and 3D classification

We encourage you to explore the things we learned today as well as the other options available in nextPYP. On day 2 we will demonstrate nextPYP’s functionality for on-the-fly data pre-processing.