Gnomesh User Manual

1. Projects

A project holds the data files you add to it, one workflow, and the list of people it is shared with. Create one from Projects → New Project in the site header, or the New Project button on the My Projects page. A project needs a name; the description is optional. A new project starts as Active and Private.

The My Projects page

The My Projects page lists the projects you own and the ones shared with you. Above the list you can filter by Status, Role, Tag and creation date, choose a sort order, and search by name, description or tag (Ctrl + K jumps to the search box). Detailed view switches the cards to a table.

Each card shows the project's name, description, your role, its status and creation date, and its actions:

• Open opens the project.
• Duplicate (the copy icon) makes a new private project named “name (Copy)”, owned by you, with the same description. It does not copy the files, the workflow, tags or collaborators. You need to be the owner or have the Copy permission.
• Delete (the trash icon, shown to the owner) deletes the project, its workflow and its files after you confirm. Only the owner can delete a project.

The project page

The top of the project page holds the project's details:

• Name and description: click either to edit it. Enter or clicking away saves; Esc puts the old text back.
• Status: Active, Inactive, Complete, Paused or Canceled.
• Visibility: Private or Public (see Sharing).
• Owner, with your role, and a Transfer button for the owner.
• Created and Last Updated dates.
• Tags: + Add Tag opens Manage Tags, where you add tags to the project, remove them, or create a new tag with a name and color. The standard tags, marked Standard, are for everyone and only an administrator changes them. A tag you create is yours: it is offered to you, and anyone who can open a project carrying it sees it there. Tags on projects you own or that are shared with you are offered too, with their owner's name beside them; hover a tag on a project to see whose it is.
• Collaborators, with + Add Collaborator for the owner.
• Export and Delete, and on a project kept in the cloud the Cloud button (see Where files are kept).

Export saves a .das archive of the project's details, configuration and list of files. It does not contain the file contents themselves. A Chromium-based browser asks where to save it; other browsers download it as they do any file.

Below the details are three cards. Input Parameters lists every File and Text node so you can give the workflow its data; the Workflow Pipeline holds the canvas; Results lists each plot and exported file from the last run, grouped by the node that made it, with its download. Each node is numbered by its place in the run order, as on its badge on the canvas.

Click the pipeline's title to fold it away. Folded, it keeps Run, Stop and Save, and shows a small map of the canvas with each node colored by how its last run went, a count of each state, and the nodes that need attention. Click the map to open the pipeline at that spot, or a listed node to open the pipeline with that node selected. Hover over a listed node to read its error or warnings. The layout button in the middle of the project header's bottom row lays the three cards out side by side, inputs on the left and results on the right, each no taller than the window, or stacks them again, whether the pipeline is folded or open.

The icons at the right of Input Parameters and Results show their items as a list, a grid of small tiles, or cards. A result's tile shows a thumbnail of the plot once the tile is on screen; a list shows none. If you can edit the project, your choice is saved with it and everyone who opens it sees the same; otherwise it lasts for the page. Your own defaults, and stacked or side by side, are under Project page in your settings.

2. Sharing and permissions

The owner shares a project by email. Press + Add Collaborator, enter the person's User Email, tick their permissions and press Add Collaborator. The person must already have a Gnomesh account and must have confirmed their email address. Access starts at once; there is no invitation to accept. The owner removes a collaborator with the Remove button on that person's row.

Only the owner can change visibility, add or remove collaborators, transfer ownership and delete the project. Transfer hands the project to another user by their (confirmed) email address; you then lose owner rights.

Public visibility lets any signed-in Gnomesh user who has the project's link view it. Public projects do not appear in other people's project lists.

3. Where files are kept

Analysis always runs in your web browser, and a project's data files are kept in this browser's private storage on this computer. Other browsers and other computers do not have them; the workflow itself is saved on the server with the project.

Files kept on the Gnomesh server are only for example projects, which an administrator prepares. Their files are fetched from the server when a workflow needs them. If a project of yours was moved to the cloud before this, its header shows a Cloud button: pressing it opens Change Processing Location, and Migrate Files downloads the files into this browser and then removes the copies on the server. If any file fails, the setting is left unchanged and the failures are listed.

4. The workflow canvas

The Workflow Pipeline section holds the canvas. A workflow is a graph of nodes. Each node does one job; connections carry its output to other nodes. A node's inputs are the dots on its left edge and its outputs the dots on its right edge. A numbered badge in a node's corner shows its place in the run order.

Above the canvas, the Input Parameters card collects the workflow's inputs in one place: each File node, with a Choose… button, and each Text node, with a text box, each marked with its output type in the color of its port. Use it to change what a workflow reads without hunting for the node.

Adding, connecting and arranging nodes

• Add a node: right-click the canvas or press Ctrl + Space. Type in Search nodes..., move with ↑ / ↓, and press Enter or click a name. The ? button shows a help panel with what the highlighted node does, what it takes in and what it gives out.
• Connect: drag from an output dot to an input dot. Gnomesh refuses a connection the receiving node cannot accept and says which kinds of data it takes. Dropping a connection on empty canvas opens the node menu showing only the nodes that can take it; the node you choose is connected for you.
• Select: click a node or a connection. Drag on empty canvas to draw a selection box. Ctrl + A selects every node; clicking empty canvas or Esc clears the selection.
• Move: drag a node. Dragging one node of a multi-node selection moves them all. With snap on, a node settles onto the grid when you let go.
• Delete: press Delete or Backspace, or use the trash button in the toolbar or the inspector. This removes the selected connection, or the selected nodes with their connections. Undo brings them back.
• Pan: drag with the middle mouse button, or hold Ctrl and drag with the left button. Zoom: the mouse wheel zooms around the pointer; double-click with the middle button to fit every node in view.

The toolbar

While Auto-Save is on (the default), the workflow is saved shortly after each change, and a one-click change such as Run After Opening or suppressing a node is saved at once. Opening a project saves nothing. A change still waiting when you leave the page is sent as the page closes; if it cannot be (Auto-Save is off, or the workflow is too large to send while closing), the browser asks before you leave.

Running a workflow

Execute runs every node whose inputs are ready, in order. A node that needs something from you, such as a file or a setting, waits and says what is missing. When a run finishes, a message gives the time taken and names any nodes that could not run or are waiting for input.

• Run one node: the inspector's Actions section has Run Node (Re-run Node once it has a result), which also runs any upstream nodes it needs.
• Leave a node out: tick Suppress on workflow runs. The node and everything downstream of it are skipped by workflow runs; running the node on its own still works.
• Leave a connection out: select it and tick Suppress this connection. It stays on the canvas, drawn dashed, but carries nothing on any run.
• When you change a node's settings, its results and everything downstream are marked out of date (“Results outdated — re-run to update”) until you run again.

Node states

A node's outline shows its state:

A suppressed node is drawn faded with a dashed outline. An “i” mark means the node has notes.

5. The inspector

Selecting a node or a connection opens the inspector panel. Drag its header to move it and its edges to resize it. The header buttons delete the selection, dock the panel to the left edge (or undock it), anchor it beside the selected node, pop it out into its own window (allow pop-ups for the site), and close it. Anchored, the panel opens next to whatever you select and follows it as you pan, zoom or move the node. Selecting several nodes shows how many are selected and their names.

Every node's inspector has, in order:

1. Name — the node's name on the canvas.
2. Notes — your own text about the node, saved with the workflow (for example, why these settings, or a warning for collaborators). It does not affect the analysis.
3. Ports — the kinds of data the node takes in and gives out.
4. The node's own settings (see the Node reference).
5. Actions — Run Node / Re-run Node, and Suppress on workflow runs.
6. Execution Results — after the node has run: its status, any message, warnings or error, and a preview of each output.

A selected connection's inspector shows where it runs from and to, its suppress setting, and a delete button.

6. Supported data

• For delimited text, the delimiter (comma, tab, semicolon or vertical bar) and whether the first row is a header are detected automatically; there are no settings for them.
• A Bruker acquisition must be added as its folder; a single file from inside it cannot be read on its own. Gnomesh keeps only the files in the folder that the reader uses.
• Bruker YEP acquisitions, and BAF acquisitions from FT-ICR instruments, are recognized but not read. The data chooser marks them “recognized, not supported”.
• The size limit is 2 GB per file. For a Bruker folder, the limit applies to each file in it.

7. Adding data: the data chooser

The Add data window opens from a File node's Choose… button in the inspector, or from that node's row in Input Parameters. It has two panes.

On this computer

Press Choose folder… and pick a folder. Click a subfolder to open it; the path above the list takes you back up. In Chromium-based browsers Gnomesh remembers the folder: next time it opens it again, or offers Reconnect to “folder” if the browser needs your permission again. In other browsers you choose the folder each time.

• Files of a type the project accepts have a checkbox; other files are grayed out and marked “not a supported type”. See Supported data for what the Parse node can read.
• A Bruker .d acquisition is one item, showing its format and how many of its files are used. Expand it to see each file, marked used or ignored. Files from inside an acquisition cannot be ticked on their own.
• When the listing has more than five items you can tick, a Select all / Deselect all button appears.

Press Add selected to copy the ticked items into the project. They appear in In this project, dimmed while they are copied, with a progress bar in the footer. A Bruker folder is copied whole or not at all.

In this project

Lists what the project holds, with a filter box. Each file shows its size; a bundle (a Bruker folder) shows “bundle”, the number of files used and its size, and expands to list them. Pick one item and press Use for this node to give it to the File node. Items you have just added are selected for you.

Drag and drop

You can drop files and folders from your computer onto a File node, the File inspector's File Selection area, or a row in Input Parameters. One file, or one acquisition folder Gnomesh can read, is added and given to the node at once. Anything else (several items, or an ordinary folder) opens the Add data window with what you dropped already ticked.

If you add a file with the same name as one already in the project, Gnomesh tells you when the contents are identical, and asks before replacing it when they differ.

8. Node reference

The data kinds below are the names shown in each inspector's Ports row.

File

Load data from a file in the project. No inputs. Gives out the file's contents: text (FileData), binary data (BinaryData), or a Bruker folder (BundleData). A Parse node almost always comes next.

• File Selection: Choose… opens the data chooser; Delete removes the chosen item from the project after you confirm; or drop a file or folder here. For a bundle, Files the parser uses from this folder lists its files.
• Detected type shows how the file will be read.
• Data Type Override: Auto-detect from extension (default), CSV / tabular text, Plain text, JSON (these three hand on text), or Binary.
• A warning appears if the node names a file the project no longer stores; choose it, or add it, again before running.

Text

Emit text you type, such as an amino-acid sequence. No inputs. Gives out TextData exactly as typed. Its one setting is Value. The most common use is a protein sequence for a Digest node.

Parse

Read raw bytes into columns or a mass-spectrometry run. Takes FileData, BinaryData or BundleData from a File node. Gives out ColumnData for delimited text, or MsData for mass-spectrometry files (see Supported data).

• Columns: once the node knows the columns, it lists them with a filter box, type chips and All / None buttons. The columns you keep are the ones offered in a downstream Plot node's column lists. At present this does not remove columns from the data itself: to choose the columns written to a file, use the Export node's Columns to include.
• For a mass-spectrometry run, the columns are per-scan values: Scan Number, Total Ion Current, MS Level, Retention Time, Base Peak m/z and Base Peak Intensity.
• Thermo .raw files are read on demand: scan details are read at once and peaks are decoded from the file when a view needs them.

Mass Spec

Inspect a mass-spectrometry run in the interactive explorer, and cut it into views.

• Inputs: an MsData run (required, upper input), and optionally peptide lists (PeptideData) from one or more Digest nodes on the lower input, to validate the run against.
• Output: the run, passed through unchanged, plus the views it generates: the chromatogram, one combined spectrum per detected chromatogram peak range (while Peaks is on in the explorer), and their MSn spectra. A Plot node downstream shows one tab per view.
• Inspector: Open MS Explorer, and a line saying how many views the last run produced. Double-clicking the node also opens the explorer. Every setting of this node lives in the explorer; see The MS Explorer.
• With a peptide list connected, peaks that match a peptide (within the label tolerance, charges 1 to 4) are labeled with the peptide sequence in the explorer's MS1 spectra.

Plot

Draw data as interactive plots. Takes ColumnData, MsData or PeptideData; several inputs may be connected, and each is plotted in turn. Gives out PlotData (the plots), which an Export node can write as images.

• From columns, each plot has one X column and one or more Y series.
• From a Mass Spec node, each view becomes its own tab, fixed to that view's data.
• From a Digest node, each plot is a sequence coverage map of the protein, and the plot type is fixed.
• Inspector: the plot's type, data and state, a thumbnail, Open plot editor… (or double-click the node) and Download PNG. Every option is in the plot editor.

Digest

Cleave a protein sequence into peptides with an enzyme or chemical agent. Takes TextData (a protein sequence); with several sequences connected, it digests each and gives out one PeptideData list per sequence, in connection order. Each peptide carries its monoisotopic neutral mass.

• Cleavage Agent (required): None (no digestion), Trypsin, Trypsin/P, Lys-C, Lys-C/P, Lys-N, Arg-C (clostripain), LysargiNase, Glu-C (V8) in ammonium bicarbonate, Glu-C (V8) in phosphate pH 7.8, Asp-N, Asp-N + N-terminal Glu, Chymotrypsin (high specificity), Chymotrypsin (low specificity), Pepsin (pH > 2), Thermolysin, Proteinase K and Elastase. A note under the list describes the chosen agent's rule.
• Missed Cleavages: 0 to 5; 2 if left empty.
• Sequence Input → Format: Auto-detect, Raw residues or FASTA. A single FASTA record only. The ambiguity codes B, Z and X are rejected because they have no exact mass.
• Peptide Filters: minimum and maximum length (residues) and mass (Da). An empty field means no limit.
• Modifications: tick any as Fixed or Variable: Carbamidomethyl (C), Carboxymethyl (C), Oxidation (M), Deamidation (N/Q), Phospho (S/T/Y), Acetyl (protein N-term), Pyro-glu from N-term Q, Pyro-glu from N-term E, Amidated (C-term). Max Variable Mods per Peptide is 1 if left empty. The digest refuses to run past 20,000 peptide forms.
• Last Digest summarizes the result: the number of peptide forms, distinct sequences and sequence coverage.

Simulation

Generate a synthetic LC-MS/MS run from digested peptides. Takes PeptideData from a Digest node (the first connected list only) and gives out MsData: MS1 survey scans and, by default, data-dependent MS2 scans. Use it to try a workflow without instrument time, or to see what a peptide set would look like.

• Ionization: Charge Model (Fixed range, or Capped by basic residues K + R), Min and Max Charge (default 2 and 3), Polarity.
• Chromatography: Retention Model (Hydropathy gradient, or None for a single scan), Retention Scale, Gradient (min), MS1 Cycle (s), Peak Width FWHM (s), Base Intensity. The gradient values must be greater than zero.
• Acquisition: MS Levels (MS1 + data-dependent MS2, or MS1 survey only), Top N, MS2 Scan Time (s), Dynamic Exclusion (s), Fragment Ion Series (a, b, c, x, y, z; b and y by default).
• Isotopes: Model (Exact, from each peptide's own formula; Averagine; or Monoisotopic only) and Peaks per Ion.

Simulated intensities carry no abundance information: every peptide gets the same amplitude, and fragment intensities are uniform. Retention times are an ordering, not a chromatographic prediction. To compare a simulated run with a real one, tick Ignore retention time on the Comparison node.

Comparison

Compare two upstream outputs and produce a match-confidence score. Two required inputs, one per side. Both must be the same kind; the method follows from it. Gives out a score between 0 and 1 (ComparisonScore).

The node gives two scores. The score above is the method's, and it is the one on the node's badge. The agreement is the share of every value the two sides hold, field by field, that agrees. Both leave out the fields you ignore and accept the error threshold you set, both from the Comparison Explorer. Differences in descriptive fields, such as the instrument or the file name, never count against either score.

The agreement compares every value of both inputs, which on two large runs takes about as long again as the score, so it is computed when the Comparison Explorer opens. Tick Compute the agreement score on every run to have each run compute it instead.

Last Result shows both scores; Open Explorer (or double-clicking the node) opens the Comparison Explorer.

Export

Package the upstream output as a downloadable file. Takes one input: ColumnData, TextData, MsData or PlotData. It produces a download rather than an output connection.

• Filename and Format; for columns, Columns to include (unticked columns are left out of the file).
• After a run, Last Export shows the file and a Download button. Five or more files download together as one .zip.
• An image's size is the plot's Export width and Export height, set in the plot editor.

Debug

Show what is flowing through a connection, and pass it on unchanged. Takes and gives out ColumnData, TextData, PeptideData, MsData, ComparisonScore or ExportData. Because it passes its input through, adding one never changes the result.

• Output Destinations: Console (the browser's developer console), Inspector (a Debug Output panel in the inspector, with a Summary or Full verbosity), and File (a Download full output button after a run, which saves the complete output as a text file).

9. The MS Explorer

Open it from a Mass Spec node: Open MS Explorer in the inspector, or double-click the node. It shows the run's MS1 data. Close it with the close button or Esc. Detection settings, hand-made ranges and peak labels are saved on the node. Looking around, or dragging a plot taller, does not make the node's result out of date; only a setting that changes what the node computes does.

A node with no current run opens too, with its settings and no data. Change them, then press Run node, and the explorer shows the data when the run finishes.

Layout

• Total Ion Chromatogram at the top, with the selected scan or scans and their retention time.
• A strip with ◀ and ▶ to step through the detected peak ranges, and the selected range's statistics.
• Mass Spectrum of the selection below, and beside it, when you open one, either the XIC or the Fragments (the buttons in the spectrum's header; press the open one again to close it, and the browser remembers the choice).
• A sidebar on the right, opened from the header: Peptides, Coverage and Labels, one tab at a time. Peptides and Coverage appear when the node matched peptides against the run.

Chromatogram peak detection

• Sensitivity (0 to 100) sets how readily peaks are found. At 0 none are found; at 50 (the default) the threshold is the knee of the score distribution; at 100 every local maximum above the floor is a peak. Every 25 steps is a factor of 10 on the threshold. Below 50 the slider also merges neighboring humps into one range, more strongly the lower it goes; higher values keep them apart. Settings saved on the old −10 to +10 slider move to this one with the same threshold (the old 0 is 50, and each old step is 5).
• The gear opens the detection settings: Min score with Auto (typing a value turns Auto off), Algorithm (Local maxima, or Derivative (Savitzky–Golay)), Fit peaks (No fit, EMG (tailed) or Gaussian, which fits each group of overlapping peaks so each reports its own area) and Padding % for Auto-size.
• Peaks turns the peak ranges on and off. It also decides whether the node produces a spectrum view per range.
• Reset view shows the whole run; Auto-size frames the detected peaks.

Peak ranges

• Hover a range for its statistics: apex, retention-time span, number of scans, height, prominence and area.
• Click a range to select it; the spectrum combines its scans. Ctrl + click a range selects only its apex scan. Click outside every range to select one scan, or drag to select a span of scans.
• ← / → (or ◀ ▶) step through the ranges.
• Adjust by hand: drag a range's edge to move it; on an edge two ranges share, Shift + drag moves both. Alt + drag draws a new range. After clicking in the chromatogram, Enter makes a range from the current selection, and Delete or Backspace removes the range you last clicked. Esc cancels an edge drag or a draw.
• Revert ranges (shown once you have changed a range) discards every hand change and rebuilds the ranges from detection.

The spectrum

• A vertical cursor line locks onto the peak a click would take, and a tooltip then gives its m/z and intensity, and, for a peak the instrument fragmented, its assigned charge and number of MS2 scans.
• Fragmented peaks are colored by assigned charge, with a legend, and marked ▾. Clicking the mark shows that peak's MSn spectrum in the Fragments pane, opening it if it is closed.
• The spectrum has its own Sensitivity and gear: Min SNR with Auto, Auto floor %, Padding %, and Combine scans — how several selected scans become one spectrum (Average, Sum or Max of points at the same m/z).
• Fast display (approximate wide selections) reduces very large selections to display resolution. Peak detection always uses the exact combined points.

Extracted ion chromatogram (XIC)

Press XIC in the spectrum's header. Then click a spectrum peak: the panel beside the spectrum shows that m/z, within the label tolerance, over the whole run, with the selected range shaded, and the spectrum marks that m/z window with a band named XIC. The XIC has its own Sensitivity and Min score; its algorithm and fit are the chromatogram's. The chromatogram above is never replaced; press XIC again to close the panel.

Fragments (MSn)

• Press Fragments in the spectrum's header, or click a fragmented peak (or its ▾) on the spectrum; with the XIC open, a peak click shows its XIC instead. The pane beside the spectrum draws the MSn spectrum of the precursor you pick, level with the spectrum, and lists the selection's fragmented peaks under it as one row of chips (precursor m/z and charge; the plot's title adds the scans and activation).
• The chips scroll sideways, with the mouse wheel too. Past five, a filter appears: type any part of a precursor's description, such as 445, hcd 4813 or z = 3. ← / → move between chips.
• The MSn spectrum marks its precursor's m/z window with a band named precursor. Once an MS2 spectrum is drawn, its own fragmented peaks (MS3) follow its chip, marked ↳; a ▾ on it picks them too.
• A spectrum is combined only when you pick it. Stepping to another range keeps the same precursor picked when that range fragmented it too.
• Dragging the spectrum's bottom edge sets the MSn plot's height too, and an exported MSn tab keeps its proportions.

Peptides and coverage

• Peptides lists every peptide the node looked for and whether it was found. Click a found one to go to the range and spectrum it was matched on. The summary's covered share opens the Coverage tab.
• Three settings above the list say how the node searches, and apply when it runs again (a Run node button appears until then): Min confidence; Estimate false matches (decoys), which also searches two sets of shifted peptides and reports how many of the matches chance alone would make, at about three times the search time (off by default); and Search at, which looks for each peptide at the apex of any peptide's ion trace, or only at its own (fewer matches, fewer of them by chance).
• Coverage draws the protein's sequence coverage map from the identifications the node's last run accepted: identified, seen below the confidence threshold, and not seen. It is the same map as the Coverage tab of a Plot node the Mass Spec node feeds. With more than one digest connected, pick which.

Peak labels

• Press Labels to open the sidebar's Labels tab. While it is open, clicking a spectrum peak adds a label (or selects the existing one) and you can type its text.
• On the plot, click a label to select it and click again to edit it in place; Enter keeps the edit and Esc cancels it. Each entry in the panel has a Remove button.
• Labels are stored by m/z, not by scan, so a label appears on any selection with a peak within the Tolerance (ppm or Da). The tolerance is always capped below the isotope spacing. Changing it also changes how scans are combined.
• Default label is the text every unlabeled detected peak shows, with tokens such as {x:.2} (m/z to two decimals).
• Show computed labels hides or shows labels the software wrote, such as peptide matches (drawn in violet).
• Import… merges labels from a JSON export. Copy from node copies another Mass Spec node's labels into this one. Export labels saves them as JSON; they are already saved with the workflow.
• The node's labels are drawn on its views in downstream Plot nodes.

Zoom, pan and export

• The mouse wheel or a pinch zooms along the horizontal axis; the vertical axis fits the visible data. Right-drag, middle-drag or a two-finger drag pans. Double-click resets the view.
• When the plots are taller than the window, scroll the explorer with the scrollbar, or Page Up / Page Down.
• Export in the header offers Export as PNG, Export spectrum data (CSV) and Export peak labels (JSON).

Changing the explorer's detection settings or ranges marks the Mass Spec node and everything downstream out of date. Run the workflow again to update the views in Plot and Export nodes. Label changes do not need a re-run.

10. The Comparison Explorer

Opens from a Comparison node that has run. The top of the window shows the node's score, its method, its agreement score and a one-line verdict on the whole comparison. Close it with the close button, Esc, or by clicking outside the window.

Data: every field of A beside B

The Data view is a tree of both inputs side by side, to any depth: a run's own fields, its scans, each scan's fields and peaks, a table's columns and cells, a text's lines. Opening a row opens it on both sides. Lists are paired by what identifies their items, not by position: scans by scan number, columns by name. One extra scan therefore marks that scan only, not every scan after it. Each row shows its value on each side and a note saying how they agree.

• Green: the two sides agree, exactly or within rounding. The note names the rounding: equal at f32 (one side was stored in single precision) or within a few f64 ulps (double-precision arithmetic). Rounding is a difference shows rounding as red instead.
• Red: the values differ, or a field is on one side only. The note gives the difference and the relative error.
• Yellow: a descriptive field differs (see below). A row whose only differences below it are descriptive stays green, and its note counts them.
• Gray: a field the node ignores.

Differences only hides the rows that agree; First difference opens the tree down to the first one. The bar under the tree shows where the selected row is on each side.

Ignoring fields

The Ignore box on a row leaves that field out of both of the node's scores. The node keeps it and runs again at once. For a field inside a list, such as a scan's, Ignore in every item in the bar under the tree ignores it in every scan. Ignored lists what the node ignores, with Stop ignoring beside each. The node keeps its ignores in the workflow, so they hold for every later run. A field ignored through a broader rule shows its box locked; stop ignoring that rule in the list.

The error threshold

Ignore rel. error ≤ sets a relative error at or below which two numbers count as equal: in the tree, its plots and both scores. The node keeps it and runs again. Set it with the slider, on a scale from 1e-16 to 2, or type a value such as 1e-9 or e-9; the left end of the slider is off. For the spectral angle, the error is of each m/z bin's share of its spectrum. A value present on one side only is never within a threshold: there is no second value to compare it with.

Descriptive fields

Some fields describe how data was acquired or stored rather than what it measured: the vendor, the instrument, the software, dates, the source file, a scan's native id. When they differ they are shown in yellow and never count against a score. Which fields are descriptive is set in two places, as a tree of every field with its type and a description:

• Admin → Descriptive Comparison Fields (administrators): the defaults for everyone. Export saves them as a file and Import loads such a file, to copy them to another server.
• Profile Settings → Descriptive Comparison Fields: your own choices, each Default, Descriptive or Data. They apply to every comparison you run, so two people's agreement scores can differ. A workflow page reads them when it opens; reload it after a change.

A field under a descriptive field is descriptive too. A node cannot change which fields are descriptive; it can only ignore fields.

Plots of a row

A row holding numbers on both sides offers plots in the bar under the tree:

• Mirror: A drawn upward, B downward; the points that differ are red, and so are the bands of the axis where they fall.
• Error: B − A at each point, against the band single-precision rounding allows. Relative error: the same, divided by the larger value.
• Heatmap: on the top row of two runs or two tables, the difference over the whole of each.

On the top row of two runs, or on their scans row, the list beside the buttons chooses whether the plots draw the TIC or the whole-run spectrum.

Score

The Score view shows what the node's method compared, where it has a view: for two runs, a heatmap of the difference over retention time and m/z; for two texts, the line diff. A table's cells are in the Data tree. Export saves the breakdown as CSV.

11. The plot editor

Open it with Open plot editor… in a Plot node's inspector, or by double-clicking the node. Every change is saved to the node as you make it; there is no separate save. Close it with the close button or Esc.

Header

• Basic / Advanced sets how many options are shown. Advanced adds more axis and legend options and All settings, the plotting engine's complete option editor.
• Presets: apply a saved preset; save the current settings as a preset; delete one; and, under the gear, start new plots from a preset or from the house style, and export or import your presets as JSON. Built-in presets: Screen, Paper — single column, Paper — double column, Presentation.
• Run runs the node and redraws; PNG downloads the current plot.
• The plot's name, beside the title, shows which plot is on screen and where it sits, such as Chromatogram › 5.98 min › 445.1203 (MS2). Press it to choose another plot.
• The size beside it, such as 1200 × 800 · 3:2, opens the export size and shape.

Plots: several plots per node

Pressing the plot's name opens the Plots list. It holds the node's own plots, then each Mass Spec node's views as a tree: the chromatogram, the ranges cut from it, and the MSn spectra under each range. Coverage maps follow. Press › beside a row to open it, and press a row to show its plot; the list closes. A view not drawn yet is gray, and is drawn when you choose it. Search by name, retention time or m/z. The arrow keys move through the list and open or close rows. The pin keeps the list open as a panel beside the plot; press it again to unpin.

+ in the list adds a plot with its own columns and settings. The plot's name, set in the selection panel when nothing is selected, is also its export file name. Remove this plot asks for a second click; a node always keeps at least one plot.

Size and shape

The size opens the export width and height, 64 to 8192 pixels, which set the exported image and the thumbnail. Choose a shape (1:1, 4:3, 3:2, 16:9) to keep it while you change the width, or Free to set both sides. The link between the two fields keeps the present shape. A Mass Spec view keeps the shape of its plot in the Mass Spec explorer, so its export looks like what you saw there: set its width here, and change its shape by dragging the plot's lower edge in the explorer. In the editor chooses whether the canvas shows the plot at that shape, with its size and scale under it, or fills the window.

Data

Choose the X column, then a Y column for each series; + Series plots another column against the same X. Each series card has its own color and a second-axis option. When the source has already run, choosing a column redraws the plot; otherwise press Run.

Settings

The Settings panel has a heading for each group; press one to open it, and open as many as you like. The search shows only the groups that hold what you type.

• Essentials: Plot type, Title, X label, Y label.
• Type & style: settings for the chosen plot type.
• Axes & ticks: for each axis, tick style (Decimal, Scientific, Engineering, Compact (SI)), decimals, target number of ticks, and scale (Auto, Fixed range, or Fit visible).
• Legend: show or hide, inside or outside the plot, position, font size, and whether clicking a legend entry hides that series.
• Colors: background, grid, origin line and text, and Show grid. Color swatches open a picker with transparency. A translucent background exports as a transparent PNG; JPEG has no transparency.
• Interaction: which axes zoom and pan, keeping the view inside the data, cursor crosshair, hover tooltip, whether scrolling over an axis zooms that axis only, and whether Ctrl + drag moves the title, labels and legend.
• Data labels & notes: drag + Data label onto a point, or + Note onto the canvas (or click either to drop one in the middle). Data label text can use tokens such as {x} and {y}.

On the canvas

• Scroll to zoom; over an axis, the scroll zooms that axis only (when that option is on). Right-drag pans, left-drag selects, double-click resets the view.
• Ctrl + drag moves the title, a label or the legend (when that option is on).
• Click the title, an axis label, a data label, a note or the legend to edit it in the selection panel; click a selected label again to edit its text in place. A legend entry selects its series, whose name, color, line width and axis you can then change. The selection panel's trash button deletes a selected label, and its reset button puts a moved element back.
• Esc deselects; pressed again, it closes the editor.

The side panels can be resized by their inner edges (double-click an edge to reset it), hidden with their chevrons, and rearranged by dragging a panel's grip. Your layout is kept with your account.

12. Exporting results

• Plots as images: connect a Plot node to an Export node and choose PNG or JPEG; it writes one image per plot tab. A Mass Spec → Plot → Export chain therefore writes the chromatogram and one spectrum per range. For one plot, the Plot inspector's Download PNG or the plot editor's PNG button is quicker.
• Tables: Parse → Export writes CSV, tab-separated text or Excel.
• MS data: an Export node connected straight to Parse or Mass Spec writes the run as MS Data JSON.
• From the MS Explorer: its Export menu saves the current plot as PNG, the spectrum data as CSV, or the peak labels as JSON.
• The workflow: Export Workflow to File in the toolbar saves a .workflow.json that Import Workflow from File can load into another project.
• The project: the project header's Export saves a .das archive of the project's details and file list (not the file contents).

13. Keyboard and mouse shortcuts

On a Mac, use Cmd where the tables say Cmd.

Workflow canvas

Projects

MS Explorer

Comparison Explorer (Data tree)

Plot editor