🧬 PRIDE — POKY Relaxation and Internal Dynamics Engine

Overview

PRIDE (POKY Relaxation and Internal Dynamics Engine) is a single, persistent GUI dialog that unifies POKY's relaxation and dynamics analyses, which previously existed only as separate example scripts run one at a time from POKY Notepad ("File > Run Python Module"). PRIDE keeps the same fitting equations and file-export conventions as those scripts, but presents every analysis as a tab in one window, with an embedded, zoomable plot and results table for each.

Each tab shares a common layout: a row of buttons to pick the relevant spectra (or paste data), a Run button, a scrollable per-residue results table on the left, and an embedded matplotlib plot with pan/zoom/save toolbar on the right. Several tabs also feed values forward to later tabs — for example, the TauC computed in the T1/T2 tab, or the tau_c computed by TRACT, can be pulled directly into the Model-Free tab with one click, instead of retyping numbers.

T1/T2, R1/R2, TauC, M.W.

This tab fits longitudinal (T1/R1) and transverse (T2/R2) relaxation decay curves from a series of spectra whose names each embed a time-delay parameter (in milliseconds), the same naming convention used by the original R1_R2_script.py and T1_T2_TauC_script.py family of scripts. Peak intensities are read out at each reference peak's position across the series and fit per residue to

• T-decay: I(t) = a · exp(−t / T)
• R-rate: I(t) = a · exp(−R · t / 1000)

Selecting T1 and T2 spectra both allows PRIDE to additionally estimate the rotational correlation time (TauC) and an approximate molecular weight, using the same NESG empirical correlation as before:

• Tc (ns) = 1 / (4π·N_freq) · √(6·T1/T2 − 7) · 1000
• M.W. (kDa) = (Tc − 0.472) / 0.58

TauC/M.W. can be computed either as a whole-protein average (with residue filters for minimum ¹H chemical shift, maximum T1/T2 fit deviation, and manually excluded residues — matching T1_T2_TauC_script3.py) or per residue (matching T1_T2_TauC_script2.py).

HetNOE

Computes the steady-state {¹H}-¹⁵N heteronuclear NOE ratio I_NOE/I_ref for every assigned peak common to a reference and a NOE-saturated spectrum, unifying hetnoe_script.py and hetnoe_script2.py. Uncertainty is propagated from spectral noise,

• σ = √[(ρ_ref/I_ref)² + (ρ_noe/I_noe)²]

with an option to use POKY's Monte-Carlo noise resampling (1000 samples) instead of the spectrum's stored noise estimate for a more conservative error bar.

SEA-TROSY

Extracts exchange rates (K_ex) from a SEA-TROSY (Solvent Exposed Amides with TROSY) buildup series (Pellechia et al., J. Am. Chem. Soc. 2001, 123, 4633), fitting each residue's peak-height buildup to

• I(t) = a · (1 − b · exp(−t / K_ex))

A minimum-deviation threshold (default 30) filters out poorly-determined fits from the plot (though every residue still appears in the results table), matching the original SEA_TROSY_script.py behavior.

CPMG Relaxation Dispersion new

The CPMG tab is new in PRIDE — none of the original misc/ scripts implemented it. It measures R_2,eff as a function of CPMG pulsing frequency (ν_CPMG) from a series of spectra recorded at a fixed total relaxation delay T against a no-CPMG reference spectrum,

• R_2,eff(ν_CPMG) = −1/T · ln(I(ν_CPMG) / I₀)

with uncertainty propagated from each spectrum's noise estimate, and fits the resulting dispersion curve per residue to one of two models:

The general model is the exact 2-state solution and reduces to the fast-exchange formula when k_ex ≫ Δω, so it is the more broadly applicable choice; the fast-exchange model needs fewer, better-determined parameters when the data are firmly in that regime. A per-residue dispersion-curve view (data + fit overlay, selectable from the results table) sits alongside the usual summary plots (k_ex, R_2,0, Φ_ex/p_B, or χ² versus residue number).

CEST new

The CEST (Chemical Exchange Saturation Transfer) tab is also new in PRIDE. It builds a normalized saturation-transfer profile I/I₀ per residue from a series of spectra whose names embed the (signed) saturation-offset frequency in Hz, against a reference spectrum recorded with the saturation field far off-resonance, and fits each profile to a full two-state Bloch-McConnell simulation rather than an approximate closed-form model — the six-component (A_x,y,z, B_x,y,z) magnetization vector is propagated through the CW saturation field for the saturation time via the matrix exponential of the exchange/relaxation/precession evolution matrix, following Trott & Palmer, J. Magn. Reson. 170, 104 (2004), and the CEST review of Vallurupalli, Bouvignies & Kay, ChemPhysChem 20, 1928 (2019).

Free fit parameters are the minor-state population p_B, the exchange rate k_ex, and the chemical shift difference Δω (ppm); R1/R2 are held fixed (with an option to also float R2) since a single CEST profile alone rarely constrains them well. Fixed R1/R2 values default to user-entered constants, or - when available - are pulled automatically per residue from the T1/T2 Relaxation tab's cached results for that same PRIDE session. A per-residue profile view (data + fit overlay, selectable from the results table) sits alongside summary plots (p_B, k_ex, Δω, or χ² versus residue number).

ZZ-Exchange new

The ZZ-Exchange tab targets slow two-site exchange - the complementary regime to CPMG/CEST above, where the two states are resolved as fully separate peaks rather than coalesced or broadened. A series of spectra with increasing longitudinal mixing time τ_mix lets magnetization leak between the two states via chemical exchange, producing two auto peaks (A→A, B→B) and two exchange cross peaks (A→B, B→A) per residue. All four intensity trajectories are fit jointly to the two-site kinetic rate matrix (not an approximate single-timepoint ratio formula), following Farrow, Zhang, Forman-Kay & Kay, J. Biomol. NMR 4, 727 (1994), and Montelione & Wagner, J. Am. Chem. Soc. 111, 3096 (1989).

Pick the mixing-time series (the first number in each spectrum's name is the mixing time in seconds); peak positions come from the longest-mixing-time spectrum's own peak list, since that's where the cross peaks are best resolved. For each residue with exactly two peaks assigned to it in that reference spectrum - the two exchanging states' auto peaks - the two cross-peak positions are computed automatically (not separately picked) by combining each auto peak's own indirect-dimension (w1) shift with the other's direct-dimension (w2) shift, then reading peak heights at all four positions across the series. Free fit parameters are the minor-state population p_B, exchange rate k_ex, and R1_a/R1_b (fit independently, not assumed equal), alongside an overall intensity scale. A per-residue buildup-curve view (all four intensities vs. mixing time, data + fit overlay) sits alongside summary plots (p_B, k_ex, R1_a, R1_b, or χ² versus residue number).

TRACT

TRACT (TROSY for Rotational Correlation Times) estimates the overall rotational correlation time tau_c directly from two cross-correlated relaxation rates (R_a, R_b) without needing a full T1/T2 series, following Robson, Dağ, Wu & Ziarek, J Biomol NMR 75, 293–302 (2021). This tab combines both original scripts — poky_tract_algebraic_analysis_script.py's closed-form solution and poky_tract_minimization_analysis_script.py's numerical minimization — and reports all five results side by side:

Bond length, ¹⁵N CSA, and CSA/N-H angle are exposed as editable fields (defaults 1.02 Å, 160 ppm, 17°) rather than hardcoded, so advanced users can adjust them per protein. A "Use algebraic TauC for Model-Free" button passes the result straight to the Model-Free tab.

TauC → M.W. Calculator

A standalone calculator (from TauC_MW_script.py) that plots a single TauC value against the same NESG reference dataset of 20 rigid, roughly spherical proteins used throughout PRIDE, and reports the corresponding approximate molecular weight. TauC can be typed in directly, or pulled in with one click from either the TRACT tab or a completed T1/T2 relaxation run.

Model-Free

The Model-Free tab is new in PRIDE — none of the original misc/ scripts implemented it. It performs a per-residue Lipari-Szabo model-free fit from R1, R2, and {¹H}- ¹⁵N NOE measured at a single field, using the standard spectral density and relaxation-rate equations of Farrow et al., Biochemistry 33, 5984 (1994), and the extended (two-timescale) treatment of Clore et al., JACS 112, 4989 (1990).

The overall rotational correlation time tau_m is held fixed (from TRACT or the T1/T2 tab) while each residue is fit independently to one of five nested models, and the best model is chosen by Akaike Information Criterion (lowest AIC wins):

Data can be gathered automatically from residues that have R1, R2, and HetNOE results in common from the other PRIDE tabs run earlier in the same session, or pasted manually as whitespace-separated rows (resnum R1 R1err R2 R2err NOE NOEerr) for data collected elsewhere. Results can be plotted as S², τ_e, R_ex, or χ² versus residue number, colored by the selected model.

RCI / S² & CSI

The RCI / S2 tab predicts a per-residue Lipari-Szabo order parameter S² directly from backbone chemical shifts alone — no relaxation experiment required — using the Random Coil Index method of Berjanskii & Wishart, JACS 127, 14970 (2005). It also classifies each residue's secondary structure (helix / strand / coil) with the Chemical Shift Index (CSI) of Wishart & Sykes, J. Biomol. NMR 4, 171 (1994), shown as color-coded points on the S² plot.

Pick a chemical shift list (the same shift lists used elsewhere in POKY) and click Run. For each residue with CA/CB/CO/N/HA/H(N) shifts available, the deviation from a (neighbor-corrected) random-coil reference shift is computed per atom, combined across atom types with empirically fitted per-combination weights, smoothed twice over a 3-residue window, and converted to S² via S² = 1 − 0.4·ln(1 + 17.7·RCI).

Results can be exported to Excel or used to color a loaded PDB structure in PyMOL by S² (low S² = flexible, colored red), the same way as the Model-Free tab.

Excel & PyMOL Export

Every tab that produces a per-residue table can export it to an Excel workbook (via openpyxl) with a ready-made scatter chart and error bars, matching the layout of the original scripts' exports. The T1/T2 tab additionally offers per-residue decay-curve export (observed vs. fitted intensity) into individual workbooks, one per residue.

The T1/T2 (R2), HetNOE, SEA-TROSY (K_ex), CPMG (k_ex), CEST (p_B), ZZ-Exchange (k_ex), Model-Free (S²), and RCI / S2 tabs can each generate a PyMOL .pml coloring script that maps the residue values onto a loaded structure — either a hard threshold (residues past a limit shown as red spheres) or a smooth red-gradient coloring — and copy the run command to the clipboard for pasting straight into PyMOL's command line.