Testing protocols good laboratory practice requirements

An important aspect is that of data and model stability. The raw data, for example, assay data, need to be updated regularly as new results become available. This process is tedious and time consuming and once a valid protocol for data acquisition and validation has been established, it is suitable for automation. The automatic procedure requires extensive data integrity checks (e.g., do all data points have valid structures and experimental results, how to handle conflicting results from different experiments) and a formalized automated normahzation process. This involves answering questions such as how to handle isomers, do experimental in vivo results have precedence over in vitro results or should all results be shown, should more confidence be put in results from Good Laboratory Practice (GLP)-studies, and so on. The models are then automatically rebuilt using the updated data and auto-updated as was mentioned in the QSAR section. A model may be promoted to use in the production system if it passes some defined validation tests, specific for each model, which was also mentioned earlier. 

Tests to establish the identity, strength, and purity of the test and control articles need not comply strictly with GLP requirements (e.g., protocol, QAU inspection requirements), but good documentation of anal5dical test results (usually in a laboratory notebook) and retention of raw data for such tests is a good practice. As the development process proceeds and the same material is used in both nonclinical and clinical studies, CGMP principles wiU apply to the production and characterization processes. 

---