Handling Human Interpreted Analytical Data Workflows for Pharmaceutical R&D Presented by Peter Russell
2011 Survey 88% of R&D organizations lack adequate systems to automatically collect data for reporting, analysis, and decision-making 1 12% 88% 1 Scientific Computing Research Study 2011
Overview - What Data? Different types of data in a lab Discrete values Reaction schemas Free-form text Documents, files, MS Office Web resources Chemical structures and formulations Chromatography Spectroscopy Images Graphs The most successful strategies are those that address needs in the context of sub-disciplines Anne E. Thessenand David J. Patterson, Data issues in life sciences, PMC (NIH/NLM), http://www.ncbi.nlm.nih.gov/pmc/articles/pmc3234430/, (November 28, 2011).
ACD/Labs sub-discipline is Human Interpreted Analytical Data
One-and-Done Human Interpreted Analytical Data Original Raw data acquisition File / folder copied to a file server
Manage Unified Information Technique and Vendor Independent
Tools for Multi-technique Elucidation and Verification
Manage Unified Information Technique and Vendor Independent CRO 1 CRO 2 CRO 3 Silent Automation
Optionally Employ Workflow Management
Typical LIMS Integration with an Analytical Lab LIMS Job Submission Project Status Close Job & advise submitter Job ID Job and Sample Description Submitter Request date Managers Submitters Lab Bench This is the findings 10
Examples of Transactional Database Systems with ACD/Labs
Workflows for Early Discovery Collection of data from Open Access and CROs and Auto-verification Access to raw data files by using a DB search of instrument meta data Automated structural verification using NMR and MS to ensure the CRO delivers what they have promised Synthetic chemistry Submission of problems by a chemist that need to be solved by an expert. E.g. Drug Screening Conformation analysis Structure Elucidation Purity Chemical Stability Chiral separation Quantitation Improving the choices for synthesis
Workflows for Early Discovery Chemist or CRO makes a compound and writes up experiment in ELN Chemist or CRO runs Spectroscopy for structural confirmation Spectra automatically stored in Spectrus DB Structure automatically added - linked by ELN No. in instrument metadata. Original Raw data acquisition File / folder copied to a file server Automated Structure Verification If structure and spectra are inconsistent, red flags are sent to Experts. Chemist submits a problem to the analytical expert Based on ELN number, the open access spectroscopy is retrieved Improved NMR predictions Pdf sent back to the analysis Expert might do additional spectroscopy and add to the DB Data worked up by Expert using ACD/Labs verification and elucidation
Workflows for Early Discovery Data collection & Archive Hyperlink to raw file Raw file zipped Auto-processed data Structure from ELN No.
Workflows for Early Discovery Sample Submission
Workflows for Early Discovery Drug Screening
Workflows for Early Discovery Drug Screening
Workflows for Early Discovery Drug Screening
Workflows for Early Discovery Drug Screening
Workflows for Early Discovery Drug Screening
Late Discovery DMPK Prediction of Metabolites Batch processing of dosed LC-MS datasets Subtraction from T=0 Identification of predicted and unpredicted metabolites Fragmentation analysis on MSn Collating metabolic pathways with structure and spectra Use of markush for partial elucidations To avoid having to elucidate more than once Speed up reporting of spectra, structure and data
Batch Processing
Summary View Biotrans Map Summary Table Study directory XIC TIC Kinetic plot
Mirrored View Meta data Mirrored MS2 spectra of Parent and metabolite with score
Metabolite View Summary MS2 and MS3 across all time points
Ions View Overlaid XICs
Development Impurity Resolution Process Chemists Method Development Specialists Toxicology Groups Stability Groups Analytical Chemists
Development Pre-formulation polymorph and salts
Extractables & Leachables
Summary 1. If I have seen this before, then the answer can be found fast 2. If I have the full chemical context, I can make an informed decision 3. If all the data is in one place, reporting is simple and quick 4. Our database systems are not an island. Integrations is often a pre-requesite