Linked Life Data

16132822

Source:http://linkedlifedata.com/resource/pubmed/id/16132822

Statements in which the resource exists as a subject.
PredicateObject
rdf:type
lifeskim:mentions
pubmed:issue
3
pubmed:dateCreated
2005-8-31
pubmed:abstractText
We present a novel automated strategy (PISTACHIO) for the probabilistic assignment of backbone and sidechain chemical shifts in proteins. The algorithm uses peak lists derived from various NMR experiments as input and provides as output ranked lists of assignments for all signals recognized in the input data as constituting spin systems. PISTACHIO was evaluated by comparing its performance with raw peak-picked data from 15 proteins ranging from 54 to 300 residues; the results were compared with those achieved by experts analyzing the same datasets by hand. As scored against the best available independent assignments for these proteins, the first-ranked PISTACHIO assignments were 80-100% correct for backbone signals and 75-95% correct for sidechain signals. The independent assignments benefited, in a number of cases, from structural data (e.g. from NOESY spectra) that were unavailable to PISTACHIO. Any number of datasets in any combination can serve as input. Thus PISTACHIO can be used as datasets are collected to ascertain the current extent of secure assignments, to identify residues with low assignment probability, and to suggest the types of additional data needed to remove ambiguities. The current implementation of PISTACHIO, which is available from a server on the Internet, supports input data from 15 standard double- and triple-resonance experiments. The software can readily accommodate additional types of experiments, including data from selectively labeled samples. The assignment probabilities can be carried forward and refined in subsequent steps leading to a structure. The performance of PISTACHIO showed no direct dependence on protein size, but correlated instead with data quality (completeness and signal-to-noise). PISTACHIO represents one component of a comprehensive probabilistic approach we are developing for the collection and analysis of protein NMR data.
pubmed:grant
pubmed:language
eng
pubmed:journal
pubmed:citationSubset
IM
pubmed:chemical
pubmed:status
MEDLINE
pubmed:month
Jul
pubmed:issn
0925-2738
pubmed:author
pubmed:issnType
Print
pubmed:volume
32
pubmed:owner
NLM
pubmed:authorsComplete
Y
pubmed:pagination
219-33
pubmed:dateRevised
2007-11-14
pubmed:meshHeading
pubmed:year
2005
pubmed:articleTitle
Probabilistic Identification of Spin Systems and their Assignments including Coil-Helix Inference as Output (PISTACHIO).
pubmed:affiliation
Biochemistry Department, National Magnetic Resonance Facility at Madison, 433, Babcock Drive, Madison, WI 53706, USA. eghbalni@nmrfam.wisc.edu
pubmed:publicationType
Journal Article, Research Support, U.S. Gov't, P.H.S., Research Support, N.I.H., Extramural