Fragment lists prediction

This is accomplished by creating the parent ion from the structure candidate and storing it in an ion list. The production rnles are applied and lead to a series of ionic strnctnre fragments that are appended to the list. In the next step, fragmentation is predicted for each ion, and a hypothetical mass spectrum is created and added to a spectrnm list. The spectrnm list contains the ion and the mass and charge value and its relative abnndance. The relative abundance is required since the same ion may occnr mnltiple times as resnlt of fragmentation of different substructures. If all ions in the ion list are compnted, the spectrum list contains all data for the hypothetical mass spectrnm of the candidate structure. 

The prediction process relies on a pre-computed set of structural fragments that give rise to toxicity alerts in case they are encountered in the structure currently drawn. These fragment lists and toxicities (e.g. mutagenicity) were drawn from the RTECS database. RTECS, the Registry of Toxic Effects of Chemical Substances, aims to list... all known toxic substances. .. and the concentrations at which. .. toxicity is known to occur currently there are over 133,000 such substance listed at http //www.ntis.gov/ products/types/databases/rtecs.asp loc=4-4-3. 

In such tables, typical chemical shifts are assigned to standard structure fragments (e.g., protons in a benzene ring). Substituents in these blocks (e.g., substituents in ortho, meta, or para positions) are assumed to make independent additive contributions to the chemical shift. These additive contributions are listed in a second series of tables. Once the tables are defined, the method is easy to implement, does not require databases, and is extremely fast. Predictions for a molecule with 50 atoms can be made in less than a second. On the other hand, it requires that the parent structure and the substituents are tabulated, and it considers no interaction... 

MSPRUNE is an extension to MSRANK that works with a list of candidate structures from CONGEN and the mass spectrum of the query molecule to predict typical fragmentations for each candidate structure. 

Our expert system is the only predictive software available based on the LSER model. This system also represents the first total codification of the rules and fragment contribution values for synthesis of the four parameter values. Previous publications have either listed partial guidelines for specific classes (27. 28) or have alluded to them (7. 13-17. 27-41). Table III demonstrates the predictive ability of the software for both the LSER parameters and contaminant toxicity. The LSER parameter values are composed from a sum of the contributions from each of the fragments, and the predicted values are generally close to the values composed by hand. For some compounds (such as acenaphthene), the values for 7T, 0, and a may not be accurately represented by a simple sum of fragment variable... 

---