Challenge open date: March 28, 2018
Challenge close date: August 1, 2018
NUS spectral reconstruction for 3D NOESY-HSQC spectra with pre-defined sample schedules
Challenge 2 will examine NUS spectra reconstructions from two uniformly sampled 3D NOESY-HSQC
spectra sub-sampled with three different sample schedule types at three different sample coverages.
Spectral reconstructions will be scored in three categories; Fidelity, a combination of
frequency and amplitude accuracy, Detection, a combination of detection of signals and
resolution, and a Combined Fidelity + Detection score.
89 amino acid protein with residues 1-56 comprising a small well folded protein fused with a linker and an IDP
107 amino acid protein, 12 kDa that is well folded, well behaved
Experiments and Sample Coverage
t1 x t2
Sampling Coverage [# of lines (% of FIDs)]
93 x 150
180 x 70
Each of the 3 coverage levels noted above is used to generate sample schedules of the following types:
This results in 9 sample schedules for each of the 2 experiments, for a total of 18 sample schedules.
These schedules are provided as examples to be used for the development of your processing script.
The scoring of this challenge will be based on a new series of sample schedules that match the
parameters stated here, but simply vary the random seed.
All spectra are located in subfolders of /NUScon/challenges/challenge_2
Each of these subfolders contains the uniformly sampled data ( US_data ) for the experiement and a directory of example sample schedules ( sample_schedules ). These sample schedules have been generated inside NMRbox utilizing a script, which is available to all contestants at /NUScon/utilities/makeNUS. All sample schedules include the t1=0, t2=0 FID and the FID collected at the maximum increment along the t1 and t2 dimensions.
A template script for this challenge is provided in NMRbox at: /NUScon/challenges/challenge_2/nuscon2018_generic.com
You should refer to the header inside the template script and the NMRbox rules for complete details on how to proceed.
Awards will be given to the “best” Fidelity score, the “best” Detection score, and the “best” combined score.
See the rules document for additional details.