NMRtist Use Cases

Written by: Piotr Klukowski
Published: July 29, 2024, 10:03 a.m.



Article content:

• Input files: which data files should I use in each use-case scenario?
• Common use-cases in the automated data analysis with NMRtist/ARTINA:
  • Peak picking
  • Chemical shift assignment
  • Chemical shift transfer
  • Chemical shift assignment (structure-based)
  • Chemical shift assignment (autocomplete partial assignment)
  • Structure determination (protein monomer)
  • Structure determination (protein-ligand complex)
• Common use-cases specific to the studied system type:
  • Defining disulphide bridge (SS-bond)
  • Defining small molecules, ligands and custom residues
• How can I improve NMRtist/ARTINA results manually?

Input files: which data files should I use in each use-case scenario?

Input file	Data types	Role in NMRtist	Example data source
Peak list	Sparky (*.list) XEASY (*.peaks)	Provides manual peak lists as input for the NMRtist applications Use-cases: peak lists are compatible with all NMRtist use-cases. REMARK: When a manual peak list is provided as application input (e.g., N15HSQC.peaks), do not include the corresponding spectrum file (e.g., N15HSQC.ucsf) as input. If both files are included (N15HSQC.peaks and N15HSQC.ucsf), NMRtist will perform automated peak picking of the spectrum, what will result in 2 peak lists being used as the application input (N15HSQC.peaks and N15HSQC_NMRtist.peaks).	Manual peak picking Manually curated peak lists generated by ARTINA
Chemical shift statistic	XEASY (*.stats)	Defines expected chemical shift values (mean, standard deviation) for the chemical shift assignment and chemical shift transfer Use-cases: chemical shift statistics are compatible with all chemical shift assignment and structure calculation use-cases. REMARK: If no chemical shift statistics are provided, NMRtist uses chemical shift statistics calculated using BMRB records. Statistics are calculated independently for each chemical shift and residue type. REMARK: Chemical shift statistics are used as soft constraints to guide optimization process. The process is still driven by experimental data, so cross-peaks must be present in the input data to support assignment. Chemical shift statistics do not need to come from experimental data or protein homologs. It is acceptable to predict them out of AlphaFold, as AlphaFold-predicted chemical shifts are typically still more informative than generic statistics (mean, std) calculated from combined BMRB records of proteins (irrespectively for their fold similarity to the studied system).	Assignment of source protein (e.g. homolog) that is used for chemical shift transfer Chemical shift predictions (e.g. UCBShift) based on an AlphaFold structure
Chemical shift list	XEASY (*.prot)	Supplies a partial chemical shift assignment for completion with ARTINA (e.g. provide manual backbone chemical shifts to complete side-chain assignment automatically with ARTINA) Provides complete assignment to be used in structure determination Use-cases: chemical shift list is compatible with all chemical shift assignment and structure calculation use-cases. REMARK: In protein-ligand complex studies is it advised to provide manual assignment of the ligand.	Incomplete list of manually assign chemical shifts (e.g. manual backbone assignment) Manually refined output of ARTINA automated chemical shift assignment Manual assignment of the ligand in protein-ligand complexes
Protein structure	PDB (*.pdb)	Defines the protein fold for the structure-based chemical shift assignment Use-cases: protein structure is compatible with all chemical shift assignment use-cases.	Protein model obtained with experimental methods (X-ray, NMR structure obtained under different measurement conditions, homologous structure) In-silico prediction (e.g. AlphaFold)
Distance restraints	CYANA (*.upl. *.lol))	Defines distance restraints to be used in the protein structure calculation Use-cases: distance restraints are compatible with all structure calculation use-cases.	Distance restraints obtained in experiments not reported in NMRtist project Prior knowledge about the system, for example information about the presence of disulfide bridge (SS-bond).
Angle restraints	TALOS (*.aco)	Specifies angle restraints used for protein structure calculation Use-cases: angle restraints are compatible with all structure calculation use-cases.	Angle restraints obtained in experiments not reported in NMRtist project

Peak picking

• Create and open NMRtist project
• Upload NMR spectra to the project storage
• Set spectra metadata, such as experiment type and axes order, in the project storage
• Run the application "ARTINA: peak picking" using the NMR spectra you would like to analyze as the input.

Chemical shift assignment

• Create and open NMRtist project
• Upload NMR spectra to the project storage
• Set spectra metadata, such as experiment type and axes order, in the project storage
• Run application "ARTINA: chemical shift assignment" using NMR spectra you would like to analyze as the input.

Chemical shift transfer

Chemical shift transfer allows finding the unknown assignment of the target protein, given a limited set of spectra of the target protein and the known assignment of the source protein (e.g. homolog).
The chemical shifts of the source and target proteins must be similar, but not necessarily identical.

• Create and open NMRtist project
• Upload NMR spectra of the target protein to the project storage
• Set spectra metadata, such as experiment type and axes order, in the project storage
• Prepare and upload *.stat file with chemical shifts and their tolerances of the source protein. Example *.stat file is available in user manual
• Run application "ARTINA: chemical shift assignment" using NMR spectra and *.stat file as the input. NMRtist will automatically recognize *.stat file in the input and trigger the chemical shift transfer.

Structure-based chemical shift assignment

The structure-based chemical shift assignment assigns chemical shifts given a set of NMR spectra and 3D structure model.
This use case is designed for NMR studies, where the fold of the protein is known. Alternatively, one can use AlphaFold prediction as the input, if experimental structure (e.g. X-ray) is not available.
An extensive evaluation of this procedure is available in our manuscript (doi/full/10.1126/sciadv.adi9323).

• Create and open NMRtist project
• Upload NMR spectra to the project storage
• Set spectra metadata, such as experiment type and axes order, in the project storage
• Upload *.pdb file with structure model to be used for structure based assignment
• Run application "ARTINA: chemical shift assignment" using NMR spectra and *.pdb file as the input. NMRtist will automatically recognize *.pdb file in the input and trigger the structure-based assignment.

Chemical shift assignment (autocomplete partial assignment)

Applicable when a partial assignment (e.g., backbone) is available from previous studies, and the objective is to find the complete assignment given newly recorded spectra (i.e., backbone and side-chain).

• Create and open NMRtist project
• Upload NMR spectra of target protein to the project storage
• Set spectra metadata, such as experiment type and axes order, in the project storage
• Prepare and upload *.prot file with known chemical shifts. Example is available in user manual
• Run application "ARTINA: chemical shift assignment" using NMR spectra and *.prot file as the input. NMRtist will automatically recognize *.prot file in the input and trigger the chemical shift completion.

Structure determination (protein monomer)

• Create and open NMRtist project
• Upload NMR spectra to the project storage
• Set spectra metadata, such as experiment type and axes order, in the project storage
• Run application "ARTINA: chemical shift assignment" with input data you would like to use for structure determination.
• Verify that the output of the automated chemical shift assignment does not contain any large anomalies. Ensure that the application output contains many strong assignments before moving to structure calculation stage.
• Run application "ARTINA: structure determination" with input data you would like to use for structure determination.

Structure determination (protein-ligand complex)

• Prepare CYANA library file with definition of the ligand
• Create NMRtist project using CYANA sequence (*.seq) and ligand definition (*.lib). Open the project.
• Upload NMR spectra to the project storage
• Set spectra metadata, such as experiment type and axes order, in the project storage. REMARK: use H, HC, HN, H1 and/or H2 tags to specify signals observable in NOESY experiments. For example in C13NOESY, one can define "H:1-101,HC:1-100" to indicate that H-dimension contains both protein frequencies (1-100) and ligand (101), whereas HC-dimension contains only protein frequencies.
• Run application "ARTINA: chemical shift assignment" with input data you would like to use for structure determination.
• Verify that the output of the automated chemical shift assignment does not contain any large anomalies. Ensure that the application output contains many strong assignments before moving to structure calculation stage.
• Run application "ARTINA: structure determination" with input data you would like to use for structure determination.

Defining disulfide bridge (SS-bond)

• Every cysteine involved in the disulfide bridge formation must be marked as CYSD in the protein sequence.
• Every disulfide bridge must be defined using 4 distance restraints specified in the *.upl file. The following example defines disulfide bridge between residue 2 and 11:

  	2 CYS  CB     11 CYS  DCB     1.00E-02  1.00E+01
  	2 CYS  SG     11 CYS  DSG     1.00E-02  1.00E+01
  	2 CYS  DSG    11 CYS  SG      1.00E-02  1.00E+01
  	2 CYS  DCB    11 CYS  CB      1.00E-02  1.00E+01
  	

  REMARK I: Effectively, it is sufficient to change the residue numbers in the above example to define arbitrary disulfide bridge in your system. DCB, DSG are dummy-atoms used by CYANA to enforce proper geometrical arrangement of atoms forming SS-bond.

  REMARK II: Arbitrary number of SS-bonds can be defined in a single *.upl file.

• Use *.upl file with SS-bond definitions as input for the structure calculation

Defining small molecules, ligands and custom residues

• Every non-standard element of the system (custom residue, small molecule, ligand) must be defined in the CYANA library file (*.lib). Definition of each non-standard element consists of its name (typically 3-4 letter code) and specification of its atom configuration. More information about building CYANA libraries is available in CYANA wiki article. Many examples of CYANA library files are stored in ATB repository
• When non-standard elements are used in NMRtist project, one needs to provide CYANA library file at the time project is created.
• When CYANA library file is provided, one can use non-standard elements in the same manner as standard residues - by using their names in the sequence.
• Non-standard elements require special handling only at the time NMRtist project is created. Each application automatically recognizes non-standard elements and their definitions.

How can I improve NMRtist/ARTINA results manually?

The results of NMRtist data analysis can be refined manually using the "human-in-the-loop" procedure.

• For peak lists:
  • Download automatically generated peak list
  • Refine peak list manually using preferred software (e.g. Sparky, CCPN)
  • Reupload file to the project and use in subsequent application run. Note: If you provide a refined peak list, use it instead of the corresponding NMR spectrum file.
• For chemical shifts:
  • Download automatically assigned chemical shifts
  • Refine the assignment manually focusing on "weak assignments" marked in light blue in the application results
  • Reupload file and use in subsequent application run.
• For structure:
  • Download output of ARTINA run
  • Are there any NOEs present in the data that are not included in the structure calculation?
  • Prepare list of manually identified NOEs (*.upl, *.lol)
  • Reupload file and use in subsequent application run.