Peak #08 · CSP Analyzer δH 8.62 · δN 119.6
HSQC Chemical Shift Perturbation Analyzer
Map a binding site from your NMR titration in minutes. Load peak lists from Sparky, POKY, CcpNmr or a spreadsheet; get weighted chemical shift changes with statistically sound thresholds, publication-ready bar charts and spectrum overlays, a global Kd fit with ligand depletion, and scripts that colour your structure in PyMOL or ChimeraX. Your data never leave your browser.
1 · Load your spectra
Two lists are enoughAdd the peak list of the free (reference) protein first, then one bound spectrum, or every point of a titration. Paste the text or open the file exported from Sparky, CcpNmr, NMRFAM-Sparky, POKY, CARA or a spreadsheet.
How to use this step
- Experiment. Choose ¹H-¹⁵N for amide HSQC/TROSY spectra or ¹H-¹³C methyl for methyl HMQC spectra. This tells the reader which columns hold which nucleus.
- Reference. The first spectrum (teal border) is the free state. Every shift change is measured from it. Use Make reference on another card to change it.
- Bound state or titration. Add one card per spectrum. For a Kd fit in step 4, enter the total ligand and protein concentration of each sample (µM); the reference has 0 ligand.
- Check the summary line under each card: it says how many assigned peaks were read and which columns were used for ¹H and the heteronucleus.
- No data at hand? Press Load example titration for a simulated 7-point titration.
Accepted peak lists
Any text table with one peak per line works. The reader finds the assignment and the two shift columns by their headers and value ranges.
Sparky / POKY (w1 = ¹⁵N, w2 = ¹H)
Assignment w1 w2
G12N-H 109.876 8.312
K13N-H 121.034 7.986
W57NE1-HE1 129.512 10.104
Spreadsheet / CSV (any column order)
residue,H,N A12,8.312,121.876 13,7.986,121.034
CcpNmr style labels such as A.12.GLY.H and three-letter labels such as Gly12 are read too. Unassigned peaks (?-?) are kept for the optional minimal-shift matching in step 2. Side-chain NH₂ and Trp indole peaks are recognised and can be included or left out.
Methyl spectra: labels like L45CD1-HD1 keep each methyl separate, so Leu/Val pairs get their own values.
2 · Chemical shift changes
Reference vs one spectrumEach assigned peak is matched between the reference and the spectrum you choose. The combined change weights the heteronucleus so it is comparable with ¹H.
How to use this step
- Compare with the bound spectrum (for a titration, normally the last point).
- Weighting. Δδ = √[½(ΔδH² + (αΔδN)²)]. Williamson (2013) recommends α = 0.14 for ¹⁵N (0.2 for glycine) and about 0.3 for ¹³C. Some groups omit the ½; choose what your field uses and report it.
- Threshold. The σ₀-corrected method (Schumann et al. 2007) computes the standard deviation (about zero) after repeatedly removing shifts above 3σ₀, so the binding site itself does not inflate the cut-off. The default flags shifts above 2σ₀: at 1σ₀ roughly a third of unaffected residues pass by chance. Mean + SD is the classic, more conservative choice. You can also type your own values.
- Disappeared peaks (in the reference but not in the bound list) usually mean intermediate exchange at the binding site. They are shown in purple in every figure; check them before you call them “no change”.
- Click a row to highlight that residue in all figures. Use ⊘ to exclude a peak you do not trust (overlap, poor shape); it is then left out of the statistics.
Settings
Minimal-shift matching pairs each reference peak with the nearest unused bound peak. It gives a lower limit for each shift and is useful when the bound spectrum is not assigned (Williamson 2013).
Significant residues
Summary
Per-residue table click a header to sort
3 · Figures
SVG & PNG for papersA per-residue bar chart and a spectrum overlay with arrows, styled for publication. Adjust the options, then download SVG (editable in Illustrator, Inkscape or PowerPoint) or a high-resolution PNG.
How to use this step
- Bar chart. Grey = below threshold, orange = significant, red = strong, purple band = peak disappeared. Dashed lines show the thresholds.
- Secondary structure (optional): type ranges like
H 5-18, E 23-29, E 34-40to draw helices and strands above the bars. - Overlay. The reference is black. With several spectra, the titration runs from blue to red. Axes follow NMR convention (ppm decreasing to the right; ¹⁵N increasing downwards).
- Click any bar or peak to highlight that residue everywhere.
- For journals, export PNG at 600 dpi, or SVG and set the final size in your figure software.
Shift change per residue
Spectrum overlay
4 · Binding curve & Kd
Optional · needs a titrationWith three or more spectra at known ligand and protein concentrations, the shift changes are fitted to a single-site binding model that accounts for ligand depletion. A global fit shares one Kd across the residues you choose.
How to use this step
- Enter the total ligand and protein concentration (µM) for every spectrum in step 1. If the protein concentration is the same throughout, fill it in below instead.
- Residues. Automatic takes the significant residues with a peak in every spectrum. You can type your own list, such as
23, 25, 51-55. - Model. Δδobs = Δδmax · [(P + L + Kd) − √((P + L + Kd)² − 4PL)] / 2P, with P and L the total concentrations. This is valid in fast exchange, where peaks move smoothly; peaks that broaden or jump (slow/intermediate exchange) should be left out.
- Reliability. Kd is well defined only if the titration goes clearly beyond Kd and Kd is not far below the protein concentration. The tool warns you in both cases.
Fit settings
Per-residue fits click a row to plot it
Global fit normalised to each residue’s Δδmax
Selected residue
5 · Export & map onto the structure
CSV · PyMOL · ChimeraXTake the results with you: a full table, scripts that colour your structure by shift change, and a methods paragraph you can adapt for your paper.
Data table
One row per residue: shifts in every spectrum, Δ¹H, ΔX, combined change and class.
Structure colouring
PyMOL: load your structure, then File → Run Script and choose the .pml file. ChimeraX: open the structure, then open csp.cxc. Residues without data are coloured light grey. The offset lets you match construct numbering to the PDB file.
Methods paragraph edit before use
Project
Your work is kept in this browser automatically. Save a project file to move it to another computer or share it with a colleague.