Peak tables

Probably the most common representation for all kind of spectra used in computerized information and expert systems is the peak table. This very simple representation consists of a table containing all (or a certain number of the most significant) peaks appearing in the spectrum. Each peak is usually described by its position and intensity, but more information (half width, multiplicity, shape type, etc.) can be added if needed. 

Such tables are very convenient for peak-by-peak search if the inverted files containing ID numbers of reference spectra are at hand. These files must be generated in advance (Fig. 4.7). 

The problem with the peak-table-representation is that the retrieved match is rather inconvenient starting point for evaluation of the experiment. A comparison between the full-curve query spectrum and the retrieved one(s), represented as the peak table(s), is almost impossible. In order to assure better comparison a link from the table representation to the original (full-curve) reference spectrum must be maintained. However, even if such link is implemented we must be aware that retrieved results obtained using ranking of peak-tables are worse compared to results obtained by comparing full-curve spectra. 

The second problem inherently associated with peak search in the inverted file of peak vs. ID numbers is the tolerance limit within such a retrieval should be carried out. If the intervals in which the peaks are inverted are broad the search will probably yield the correct answer but the list of produced matches will be rather 

---

Neural networks have been applied to IR spectrum interpreting systems in many variations and applications. Anand [108] introduced a neural network approach to analyze the presence of amino acids in protein molecules with a reliability of nearly 90%. Robb and Munk [109] used a linear neural network model for interpreting IR spectra for routine analysis purposes, with a similar performance. Ehrentreich et al. [110] used a counterpropagation network based on a strategy of Novic and Zupan [111] to model the correlation of structures and IR spectra. Penchev and co-workers [112] compared three types of spectral features derived from IR peak tables for their ability to be used in automatic classification of IR spectra. 

Figure 1. Flow chart of the Polymer Analysis program. The program Is entered from a larger program, NMRl. A database must be chosen or created for the spectrum at hand and a statistical model chosen. Options In the main menu Include calculation of probabilities associated with the model, simulation of spectra, and modification of the peak table or database.

Last, a variety of options are available to modify the peak table or polymer database. The former include deleting and inserting peaks, calibration, and axis unit conversion. 

Reasonable NO conversion can be achieved using n-decane as reductant. In the absence of sulfur dioxide, the catalytic activity is roughly related to the r ucibility of the Cu phase of Cu ions in zeolites the reaction temperature needed to reach 20% NO conversion parallels that of the TPR peak (Table 7). This relation also practically holds for Cu on simple oxides, therefore a redox mechanism in which reduction of Cu + cations is the slow step could account for the results. 

With a peak table based method, the biomarkers are extracted from an unknown spectrum, and those biomarkers are then used to compare the unknown to a reference signature. 

Peak table based methods first extract biomarker peaks from a spectrum and then use those peaks for organism identification. These methods have the ability to ignore peaks that are due to background or extraneous factors. 

If, when a given organism A is present in the sample, biomarker peak i appears with probability ph Let xt = 0 if biomarker peak i is not observed in the unknown sample, and xt = 1 if biomarker peak i is observed in the unknown sample. Then the likelihood of the observed peak table is given by... 

If we let H0 represent the hypothesis that organism A IS NOT present in the sample, and HA represent the hypothesis that organism A IS in the sample, then the likelihood ratio for H0 versus HA is given by the probability of the observed peak table under HA divided by the probability of observing the outcome under H0. Specifically,... 

The NMR spectrum of the copolymer prepared from an equimolar mixture of the monomers is shown in Figure 10. In this spectrum, five well separated regions of NMR peaks were observed. The assignments of the peaks (Table III) were made by using the existing spectral information on homopolymers of 1-hexene and 5-methyl-1,4-hexadiene as well as the intensity variations among the copolymers with different monomer charge ratios. 

In addition to the above facilities which enable the analyst to save a considerable amount of time and to improve the quality of spectra, there is also the ability to store thousands of spectra on disk in a library of peak tables. Each table will consist of the wavenumbers of twenty or thirty of the most significant peaks in the spectrum together with the corresponding peak transmittance values. Several thousand tables can be stored on a single floppy disk and library searches can be conducted in a matter of seconds. After recording the spectrum of an unknown sample, a preliminary search to indicate possible structural features can be initiated. This may be followed by a complete search in which the peak table for the unknown is matched with as many library tables as the analyst has available. The computer then displays a list of ten to fifteen possible compounds in order of closeness of match using a graded scale, e.g. 0 to 9. 

High accuracy of the measurements is of utmost importance. There are 20 natural amino acids that, except for the Leu/Ile pair, have different molecular weights. However, the some of them or of some of their combinations have very similar molecular weights. The case of combinations of amino acids with similar molecular weights is especially important when the spectrum lacks a full string of ion peaks (Table 6.4). 

The weight-average branching functionality (fw) for the major MA peak (Table 11) falls between 16 and 18 for all the samples. This probably reflects the constancy of experimental preparation conditions temperature, concentration, solvent, and the ratio [DVB]/[RLi]. 

Calculation of the molecular weights of the -butyl esters of acylcarnitines, expressed as (M+H)+ molecular ions, allows the designation of each peak (Table 3.2.1). Quantitation software provided by the instrument vendor (i.e., Chemoview for SCIEX instruments) allows calculation of quantitative and semiquantitative concentrations for these acylcarnitines by comparison of the detected abundance of each acylcarnitine versus that of a designated internal standard with a known concentration. The results for each sample are further compared to age-appropriate reference ranges (Table 3.2.2). 

Solution The mlz 228 peak has 12.0% of the height of the mlz 227 peak. Table 22-2 tells us that n carbon atoms will contribute n X 1.08% intensity at mlz 228 from 13C. Contributions from 2H and 170 are small. 15N makes a larger contribution, but there are probably few atoms of nitrogen in the compound. Our first guess is... 

The hydrides HM(PF3)4, M = Co, Rh, Ir, possess a structure simUar to that of HCo(CO)4. In C3v skeletal symmetry the filled metal orbitals are of symmetry e(2), the Rh-H a bond transforms as a, and the metal-phosphorus a bonds span the irreducible representations a,(2) + e. Three low-energy peaks (Table XXIX) (169, 227) have been detected in the UPS of HCo(PF3)4, and overlapping of ionization occurs with the Rh and Ir compounds (Fig. 28). While the assignments cannot be regarded as definitive at the present time, the first two peaks in the UPS of HCo(PF3)4 probably correspond to the two 2E ionic states of predominant metal character. 

It is worth noting that the relaxation observed from dielectric measurements (Fig. 77) occurs at the same temperature as the one from dynamic mechanical measurements. In addition, the activation energies derived from the maximum of the mechanical /3 peak (Table 7) are close to those obtained from dielectric relaxation (Table 6). 

---