Skip to content

Core concepts

makeshift centers on three objects: NMRStarEntry, ChemicalShifts, and PeakList. Understand how they connect and what each does, and the rest of the API follows naturally.

NMRStarEntry

NMRStarEntry is your gateway to a BMRB deposition. It wraps a downloaded or local NMR-STAR file (.str) and exposes everything inside it as tidy tables — no parsing or format knowledge required.

entry = ms.NMRStarEntry.from_bmrb(25013)

From here you can pull sequences (what proteins/entities the entry describes), samples (what conditions they were measured under), spectrometers (the hardware), citations, cross-references (to PDB / AlphaFold), and any measured data — chemical shifts, relaxation, order parameters, spectral densities. See Reading BMRB entries for the full menu.

If NMRStarEntry doesn't have a dedicated method for something you need, two escape hatches let you reach any field:

  • entry.categories() — browse what saveframe categories the entry contains
  • entry.data_loop(category, loop_name, tags=None) — extract any tabular loop as a DataFrame

ChemicalShifts

ChemicalShifts is a tidy table of assigned backbone shifts — one row per atom. You build it from an NMRStarEntry or fetch it directly from the BMRB:

cs = ms.ChemicalShifts.from_bmrb(5363)
# or
entry = ms.NMRStarEntry.from_bmrb(5363)
cs = ms.ChemicalShifts.from_entry(entry)

cs.data is a DataFrame with columns Seq_ID, Comp_ID, Atom_ID, Atom_type, and Val (the shift in ppm). It's the standard input for most downstream analyses.

Two key operations live here:

  • Re-referencing — correct mis-referenced shifts with cs.reref("lacs") or "panav". See Re-referencing.
  • Building peak lists — go straight from shifts to assigned peaks with cs.peaklist().

PeakList

PeakList is an assigned peak table — usually an amide HSQC (¹H–¹⁵N correlations), but any dimension pair you want. Build it from shifts or read it from a CSV:

peaks = cs.peaklist()                  # from ChemicalShifts
peaks = ms.PeakList.from_bmrb(5363)    # or direct from BMRB
peaks = ms.PeakList.from_csv("peaks.csv")

peaks.data is a DataFrame with one row per peak, columns for per-dimension ppm values (e.g. H_ppm, N_ppm), and assignment labels. From here you can:

  • Compare peaks — use makeshift.spectra.map_peaklists to align experimental and reference peaks, for instance in a titration or CSP analysis. See Spectra.
  • Plot assignments — the makeshift.spectra plotting helpers take a PeakList and draw it on a spectrum.
  • Summarise completeness — peaks.assignment_string() renders a compact per-residue label string ('A' assigned, '.' missing, 'P' proline).

The workflow

import makeshift as ms

# 1. Get an entry and explore it
entry = ms.NMRStarEntry.from_bmrb(25013)
entry.datasets()       # what's in it
entry.sequences()      # the proteins

# 2. Extract chemical shifts, optionally re-reference
cs = ms.ChemicalShifts.from_entry(entry, reref="lacs")

# 3. Build a peak list
peaks = cs.peaklist()

# 4. Go further (relax dynamics, CPMG, structure prediction, etc.)
from makeshift.relaxation import RelaxationProfile
prof = RelaxationProfile.from_entry(entry)
prof.add_rigid_prediction()
prof.plot("R2_R1")

Each object is independent — you can use just NMRStarEntry for metadata, or skip straight to ChemicalShifts.from_bmrb if shifts are all you need.