Application of neural networks to the analysis of pyrolysis mass spectra.

Pyrolysis mass spectrometry is a data rich analysis technique now becoming widely applied in microbiology. Data analysis is a key step in the exploitation of the technique and the application of neural network analysis to pyrolysis mass spectrometric data offers new opportunities for classification, identification and inter-strain comparison of microorganisms in biotechnology and clinical microbiology. The use of a supervised neural network for the identification of members of a streptomycete species-group is described.

Data verification in the residue laboratory.

Residue analysis frequently presents a challenge to the quality assurance (QA) auditor due to the sheer volume of data to be audited. In the face of multiple boxes of raw data, some process must be defined that assures the scientist and the QA auditor of the quality and integrity of the data. A program that ensures that complete and appropriate verification of data before it reaches the Quality Assurance Unit (QAU) is presented. The "Guidelines for Peer Review of Data" were formulated by the Residue Analysis Business Center at Ricerca, Inc. to accommodate efficient use of review time and to define any uncertainties concerning what are acceptable data. The core of this program centers around five elements: Study initiation (definitional) meetings, calculations, verification, approval, and the use of a verification checklist.