PROTINFO-NMR documentation

• Calculations times
• Submission tips
• Troubleshooting tips
• Secondary and tertiary structure inputs and outputs
• Version information

Calculation times

The methods are usually executed on a 64 CPU cluster. Our goal is to ensure that that the prediction time for each sequence is less than 24 hours, but this of course depends on how many people submit sequences in a given day.

Submission tips

• Sequences must be specified in a single line using the one-letter amino acid notation.
• Submit sequences split up into domains if you have a good feel for where the boundaries are. This is because the complexities of most calculations are generally exponentially proportional to the lengths of the sequences, and most prediction methods are calibrated to work on domains. The programs currently perform a limited amount of automatic domain parsing, which will be enhanced in the future.
• Very short sequences are not a good idea since predictions made aren't usually reliable.
• Very long sequences are not a good idea since the computation time required is much longer. People submitting long sequences may have to wait much longer on average before their sequence is pulled for processing, compared to people submitting short sequences.

Troubleshooting tips

• Read the warnings and the error messages generated (if any) and please contact us if you think something has gone terribly wrong.

Methods

All publications on the research are accessible from our ongoing areas of research page, and all or most of the software is accessible from our software distribution server.

General notes

Following the CASP convention, up to five models may be returned (in CASP format).

Secondary structure assignment using PsiCSI

This method (published in Protein Science) uses neural networks to translate NMR chemical shifts into secondary structure information (somewhat similar to CSI) and combines it with sequence based predictions (à la Psipred). It has a sustained three-state average accuracy of 89% on a rigourously jack-knifed test set of 92 proteins for which NMR chemical shift information was publicly available. This is done before the hybrid NMR-de novo simulations are run to ensure more accurate secondary structure predictions (which helps in our simulations).

PsiCSI chemical shifts must be supplied in NMR-Star format (tools for converting to this format from other popular formats are available).

Tertiary structure determination using PROTINFO-NMR

The chemical shifts must be supplied in NMR-Star format (tools for converting to this format from other popular formats are available). This method uses an iterative hybrid NMR-de novo protocol for tertiary structure predictions.

15N and 13C edited NOESY data must be supplied as two separate files, with at least one of them required. The first three lines should contain the following:

1. Heteronucleus (15N or 13C)
2. Resolution in ppM for proton, heteronuclear correlated proton and heteronuclear dimensions
3. Water shift ppM

The rest are the list of peaks each specified in a single line as:

<peak-number> <proton-shift> <heteronuclear-correlated-proton-shift> <heteronuclear-shift>

For example:

15N
0.02 0.04 0.7
4.73
1 1.33 8.01 121.0
2 4.33 8.01 121.0
.
.
.

An optional set of one or more conformation files in the standard PDB format for NMR files (each model separated by "MODEL N" record where N is the number of the model) may also be input. These files will be used by the Bayesian constraint assignment protocol for constructing probabilities. This enables constraint assignment and structure generation in iterative cycles (using our own or other standalone software such as CYANA or XPLOR-NIH).

Version information

PROTINFO-NMR v0.1
RAMP v0.4H