Using ChopChopMF

How to use it:

1. Enter the UniProt ID of a human protein, e.g. P38398 (BRCA1).
2. Check/adjust the download folder — it defaults to Downloads and is remembered between ChimeraX sessions.
3. Click ChopChop Missense PDB. ChopChopMF downloads the matching structure and AlphaMissense scores, opens the structure in ChimeraX, and colors it by the scoring scheme shown below.

Already have your own TSV?

1. Tick Use uploaded AlphaMissense TSV file.
2. Click Browse and select your .tsv file.
3. Select the open ChimeraX model to use for chain-length detection (click ↻ Refresh model list if it's not listed yet).
4. Click ChopChop Missense PDB as above.

Hegelab

• Fetch Open the database structure with the most similar sequence. Switch Sequence to UniProt identifer, to use UniProt ID

• Search Find similar sequences in the AlphaFold database using BLAST

• Predict Compute a new structure using AlphaFold on Google servers.

How to use it:

1. Open or fetch an AlphaFold structure (via the tab above, or any model with B-factor = pLDDT).
2. Select it in the Select model to color dropdown, e.g. #1 AF-P04637-F1.
3. Click Color selected model by AlphaFold2 pLDDT. The structure is colored per-residue using the confidence scale shown above the button (dark orange = very low, blue = very high).

How to use it:

1. Select the AlphaFold structure from the list, e.g. #1 AF-P04637-F1, and click Use Selected Model — ChopChopMF fills in the UniProt ID and chain for you.
2. If your model wasn't fetched from the AlphaFold database, type the UniProt ID yourself (e.g. P04637) and pick the chain manually.
3. Click 4. Fetch UniProt Annotation & Associate to pull the UniProt annotation and link it to that chain.

Links to UniProt and AlphaFold Protein Structure Database

AlphaSync

How to use it:

1. Go to the Structure Selection sub-tab, select an open AlphaFold structure (e.g. #1 AF-P04637-F1) and click Use Selected Model — or just type a UniProt ID (e.g. P04637) directly into the field above.
2. Click ChopChop AlphaSync Residue Data.
3. Open the Residue Data sub-tab to see the per-residue table (pLDDT, SASA, RSA, surface/core, disorder, secondary structure, contacts).
4. Not sure what a column means? Check the Explanation sub-tab — click any parameter name to expand its definition.
5. Want SASA/RSA/surface/disorder/secondary-structure in Investigate too, or usable with color byattribute/select? Back in Structure Selection, click Attach fetched values to this structure. ChopChopMF first checks that the UniProt sequence position actually matches this structure's own residue numbering (comparing each position's amino acid, chain by chain) before writing anything — if it doesn't match (e.g. a cropped structure, or the wrong UniProt ID), nothing is attached and you'll see exactly why.

How to use it:

1. Select the model and chain you want to crop, e.g. Model 1, chain A. Click ↻ Refresh model list and chains if it's not listed.
2. In Residue range to keep, enter the residues you want to keep — everything else is removed. Example: 1-120,150-200 keeps residues 1 through 120 and 150 through 200, and deletes everything in between and after.
3. (Optional but recommended) Click Hide Deletion Preview first — this hides the residues that would be removed without deleting anything, so you can check the range is right. Click it again ("Reset Preview") to undo the preview.
4. Click ChopChop Crop to actually delete the residues outside your range.

How to use it:

1. Select the model and chain you want to remove entirely, e.g. Model 1, chain B.
2. Click Delete Chain.

How to use it:

1. Select the model to duplicate, e.g. Model 1.
2. Select a chain, or leave it on *(All chains) to duplicate the whole model.
3. (Optional) Tick Apply offset to duplicate and set ΔX/ΔY/ΔZ, e.g. 50 Å in ΔX, to place the copy visibly next to the original instead of directly on top of it.
4. Click ChopChop Double.

How to use it:

1. Select the map/volume you want to center on, e.g. #2 (my_map.mrc). Click ↻ Refresh maps if it isn't listed yet.
2. Click ChopChop Measure Center. The XYZ center coordinates are printed to the ChimeraX Log — keep the Log window open so you can copy them in the next step.

How to use it:

1. Ran Measure Center above? Copy the XYZ coordinates from the Log, paste them into Paste XYZ for center (e.g. 353.79, 353.79, 333.95) and click Apply. Otherwise, type the Center X/Y/Z values manually.
2. Select the structure model to copy, e.g. Model 1.
3. Select the symmetry group, e.g. C2 for a 2-fold symmetric assembly, or D3 for a dihedral trimer-of-dimers.
4. Click ChopChop Symmetry Copies.

How to use it:

1. Open your predicted complex (e.g. an AlphaFold-Multimer .cif) — it must be the only open model.
2. Click Load .json file and select the matching PAE .json from the same prediction (AlphaFold DB, AlphaFold3, or ColabFold format are all supported). ChopChopMF links the file directly to the open model and confirms it was loaded successfully.
3. Click ↻ Refresh model list, then select the two chains you want to check for contacts, e.g. chain A and chain B.
4. Set the contact distance cutoff — 5 Å is a good starting point; avoid going above 8 Å, since protein-protein interactions further apart than that are unlikely to be real contacts.
5. Optionally tick Limit to residue pairs with PAE ≤ and set a value — 12 Å is the default and a common convention for a confident contact (a distance of 8 Å combined with a max PAE of 12 Å is a widely used definition of a high-confidence interface).
6. Click ChopChop PAE. ChopChopMF draws pseudobonds between residue pairs that pass the distance (and, if enabled, PAE) cutoff.

A quick, color-coded verdict on the chain pair selected in Tab 1: is this interaction actually worth a closer look?

How to use it:

1. Select two chains in Tab 1 (a loaded .json file is only needed for the LIS/cLIS/iLIS row, not for the others).
2. Click ↻ Refresh Scores. Each score is shown with its own Select button, so you can jump straight from a number to the actual residues behind it in the 3D view:
   • pDockQ with a colored badge (Bryant, Pozzati & Elofsson, 2022): green = high confidence (> 0.5), amber = weak/medium (0.23–0.5), red = poor (< 0.23). Select contacts selects and highlights the contact residues as sticks. Note: pDockQ is derived from contact count and pLDDT only — it does not read the PAE matrix, despite living in this PAE-focused tool.
   • Buried area, computed with ChimeraX's own built-in measure buriedarea — shown as a plain number, since there's no universal "good/bad" cutoff for it in the literature. Also contact-geometry-based, not PAE-based.
   • H-bonds, computed with ChimeraX's own built-in hbonds. Select H-bonds selects them and colors each one by its own length — green = shorter/stronger, red = longer/weaker within the found set. Also not PAE-based.
   • LIS / cLIS / iLIS (Kim et al., 2024) once a PAE .json is loaded, with iLIS ≥ 0.223 badged as a high-confidence interaction. Select confident pairs selects only the residues involved in a cLIS-qualifying pair (PAE ≤ 12 Å and Cβ–Cβ distance ≤ 8 Å) — the residues actually responsible for a high iLIS.
   • ipSAE (d0chn) (Dunbrack, github.com/DunbrackLab/IPSAE) once a PAE .json is loaded — shown as a plain number, since there's no established confidence threshold specifically for this fixed chain-length variant (only the adaptive full ipSAE metric has one, and it isn't implemented here).
   • Contact PAE, once a PAE .json is loaded — the actual PAE value (Å), not a derived score: for each contact residue, the real PAE reading to its partner chain averaged over its contacts, using the same direction ChimeraX itself uses to color the PAE contact pseudobonds in Tab 1. Shown as the average over all contact residues, colored on the same blue-to-white scale as the PAE Matrix plot. Roughly ≤12 Å is confident (Tab 1's own default contact cutoff), values near/above 20 Å are uncertain. The per-residue values (not just this average) are recorded on each contact residue and show up individually, color-coded, in Investigate.
3. Click Deselect to clear the selection and reset any highlighted residues back to a neutral color, Open residue table to inspect the current selection (chain, residue number, name, pLDDT) in a sortable, exportable popup, or Export scores as CSV… to save the aggregate values for a spreadsheet or lab notebook.

Switching between Select buttons (or different chain pairs) always resets the previous highlight first, so only the residues behind the metric you just clicked are ever shown.

See Acknowledgements for the full citations of every score.

A quick visual check of the model's confidence, without leaving ChopChopMF.

How to use it:

1. Open your model (and, for the PAE-based plots, load its .json file in Tab 1 first).
2. Two sections are available: pLDDT per Residue (from the structure directly — works even without a loaded PAE file) and ipTM / pTM (only shown if present in the loaded .json — common for ColabFold, usually not available for AlphaFold3, whose confidence values live in a separate summary file), each bar colored green/amber/red for >0.8 / 0.6–0.8 / <0.6, the same traffic-light convention as the AlphaFold Guide's Confidence Metrics page. The PAE Matrix section shows the full error heatmap, with chain boundaries marked.
3. Click Open Figure under any of them to draw the current data in its own larger, freely resizable window (drawn fresh each time, so it's always up to date). That window has a Save… button to export the plot as a PNG, PDF, or SVG file.

You saw some interesting or promising results with ChopChop PAE? Now you would like to see the side chains of the pseudobonds with a good (blue) score?

How to use it:

1. Run ChopChop PAE in the first tab first, so a "PAE Contacts" pseudobond model exists.
2. Click ChopChop PAE interaction Residues. The contact residues are selected, shown as sticks, and colored by chain/heteroatom for a closer look.
3. By default the "PAE Contacts" pseudobond model is deleted afterwards to keep the scene tidy. Untick Remove 'PAE Contacts' model after selecting first if you want to keep inspecting or coloring the pseudobonds themselves.

A much more precise analysis of the PAE can be performed outside the ChimeraX environment with the PAE Viewer

Using the ChopChopMF Interface

Once you have your XML file, use the Interface Scoring tab as follows:

1. Select the target model: pick the model the XML residues belong to, e.g. Model 1. Click ↻ Refresh model list if it isn't listed.
2. Load Data: Click Select PDBePISA XML File and upload the file you just downloaded.
3. Map Interfaces: Loading the file already selects and colors the interface residues in darkorange. To (re-)apply the full 3-way scoring, click ChopChop PISA Interfaces and select the _output.defattr file - by default written next to your XML, or in the folder set centrally in ChopChopMF's Setup toolbar tool (applies to this tab and the ΔG Filter tab below; the label here shows the current value but isn't editable in-tab anymore - see 7. Setup).

4. Scoring Scheme: The tool automatically categorizes residues based on:

   • ■ Buried: Residues hidden in the interface.
   • ■ Hydrogen Bond: Specific polar interactions.
   • ■ Salt Bridge: Electrostatic interactions between charged side chains.

5. Update Visuals (optional): Type new color names (e.g., red, gold, blue) in the text fields and click Apply New Color Scheme.

1. Setting up the Filter

1. Select the target model: pick the model the XML residues belong to, e.g. Model 1.
2. Load XML: Click Load PDBePISA XML File to bring in your data. You can load more than one XML (e.g. for several interfaces of the same complex) and switch between them with the Active XML file dropdown.
3. Append Mode (optional): Toggle this to accumulate residues across all loaded XMLs at once instead of just the active one.
4. Neutral band (optional): Leave Neutral band (±ε) checked to keep near-zero residues (default ε = 0.01 kcal/mol) colored lightgrey as a baseline, so only meaningfully stabilizing/destabilizing residues stand out.
5. ΔG Cutoff: Use the slider to set a threshold, e.g. 0.50 kcal/mol. With Only show residues ≥ cutoff checked, residues below that are excluded from coloring.

2. Using the ΔG Coloring Interface

This module maps energy values to a specific color palette to highlight "hotspots" in the interface:

1. Click ChopChop ΔG Coloring to apply the energy-based colors to your structure in ChimeraX.
2. Click Plot ΔG Values to open a bar chart, scatter plot, and value list of every colored residue — handy for picking a good cutoff.
3. (Optional) Adjust the color fields under ΔG Palette and click Apply New Color Scheme to recolor with your own palette; the plot follows the same colors.

1. Pick a Model and click ↻ Refresh model list if needed.
2. Ctrl+click an atom in the 3D view to select and inspect its residue (ChimeraX's own selection shortcut - a plain click just rotates the view; Ctrl+Shift+click adds to the selection, but Investigate only shows a single selected residue at a time). You can also use another tool's Select button instead (e.g. PDBePISA's interface coloring or PAE Analysis's "Select contacts") - Investigate follows any selection, from any source. The highlighted blue bar shows which residue is currently shown. Investigate shows the residue's pLDDT (color-coded, from the structure itself) plus every custom value recorded for it by any tool (e.g. PDBePISA's residue_score, ChopMissense's MissenseScores) - PAE Analysis's pDockQ and iLIS values are color-coded here too, the same green/amber/red badge as PAE Analysis's own Scores tab, and its "PAE (Å)" value (the real per-residue PAE reading, not a derived score) is colored on the PAE Matrix's own blue-to-white scale.
3. Type a note and click Save note to attach your own free-text observation to that residue (one note per residue - saving again replaces it, clearing the field and saving removes it).

A full residue table only really works in a big window, so this tab is just a launcher - click Open Chart for every residue of the selected model at once, with:

• Its own Model selector at the top - independent of the Residue tab's, so you can flip the Chart between any open model without disturbing whatever's selected in the background window, useful for comparing two models side by side (open the Chart twice, pick a different model in each).
• A search box that filters the table to residue numbers containing what you type (e.g. 117), live as you type - clear it to see every residue again.
• pLDDT, always shown and color-coded (red/orange/yellow/blue, the same confidence bands ChimeraX itself uses) - straight from the structure, no other tool needs to have run first. Interface pDockQ and PAE iLIS are color-coded too, wherever set (same citable green/amber/red badge as PAE Analysis's Scores tab), and PAE (Å) is color-coded on the same continuous blue-to-white scale as the PAE Matrix plot. Only columns named PAE … actually read the PAE matrix itself (LIS/cLIS/iLIS, ipSAE, the cLIS contact flag, and PAE (Å)) - Interface pDockQ/Buried area/H-bonds/contact are contact-geometry- and pLDDT-based and don't depend on PAE at all, despite living in the same tool. All of pDockQ/Buried area/H-bonds/LIS/cLIS/iLIS/ipSAE are a single value for the whole interface, not per-residue math - they only appear on the actual contact residues (the same ones "Select contacts" highlights), left blank everywhere else, so the same number doesn't misleadingly show up on every residue of both chains.
• A column for every value another ChopChopMF tool can produce - PDBePISA class, PDBePISA ΔG, AlphaMissense, Interface contact, PAE cLIS, PAE (Å), AlphaSync SASA/RSA/Surface/Disorder/Sec. Str. - always shown, even before that tool has been run (blank until then), so the table itself tells you which analyses are still worth doing for this structure. Any other custom value picked up automatically gets its own column too.
• An editable Notes column (double-click a cell, type, press Enter to save - every other column is read-only). Editing a note here or in the Residue tab updates the same underlying note.
• ↻ Refresh, Export as CSV… (this model only, a one-way spreadsheet export), Export as CSV… (all models) (every residue of every open model in one file, using only the fixed known-tool columns so rows from different models always line up), and Export for AI analysis… (asks what you want the AI to look at or answer, then writes a Markdown file with that question up front, references to this guide, the AlphaFold Guide, and the ChimeraX FigureStyle guide - with an explicit instruction to actually read them, not just note them - a short explanation of what each column means, and your notes - ready to paste straight into Claude, ChatGPT, or Gemini; nothing is sent anywhere by ChopChopMF itself). To snapshot or reload the underlying .chopchop.json itself, use ChopChopMF's Setup toolbar tool (see 7. Setup).
• Every value shown here - from any tool - is also durably saved into that same .chopchop.json file each time the Chart is opened or refreshed, so it's still there the next time you reopen this structure, even in a brand new ChimeraX session. See Input & Output Files for the full picture.

Two visual ways to spot interesting residues without reading every row of the Chart yourself.

• Attribute Coverage - a heatmap, one row per residue and one column per value any ChopChopMF tool can produce (pLDDT plus the same fixed columns as the Chart). Gray means that tool hasn't recorded anything for this residue yet; teal means it has, colored by how strong that value is relative to the rest of the structure. A residue with several teal columns in a row has evidence from multiple tools at once - exactly the residues worth a closer look in the Residue tab.
• Compare Two Values - pick any two numeric values (e.g. pLDDT vs. AlphaMissense) from the dropdowns and click Open Figure for a scatter plot, one dot per residue that has both values. Useful for spotting residues that are unusual on both axes at once (e.g. low pLDDT and high pathogenicity) - something a table of numbers doesn't show at a glance. Click any point to select that residue - jumps straight to its Residue-tab dossier.

Both figures open in their own floating window with a Save… button (PNG/PDF/SVG), same as PAE Analysis's Plots tab.