How to format your data for importing into SIMCA-P - Umetrics

How to format your data
for importing into SIMCA-P
Head Quarter:
Umetrics AB
Box 7960
SE-90719 Umeå
Sweden
Phone: +46 (0)90 184800
Fax: +46 (0)90 184899
Email: info.se@umetrics.com
Content
General points ........................................................ 2
1. QSAR ................................................................. 2
2. Genomics / Proteomics / Metabonomics .............. 2
3. Spectroscopic data ............................................. 2
4. Process data (or biological time courses) ............. 3
5. LC-MS or GC-MS data ........................................ 4
6. Chromatography & Electrophoresis data ............... 4
Appendix:
GUIdInG THE WAY – USInG SIMCA-P
Detailed batch data description ............................... 5
www.umetrics.com
European Sales Offices:
Umetrics AB
Stortorget 21
SE-211 34 Malmö
Sweden
Phone: +46 (0)40 6642580
Fax: +46 (0)40 6642585
Email: info.se@umetrics.com
Umetrics UK Ltd.
Woodside House,
Winkfield, Windsor
Berkshire, SL4 2DX, UK
Phone: +44 (0)1344 885605
Fax: +44 (0)1344 885410
Email: info.uk@umetrics.com
The Standard in Multivariate Data Analysis
SIMCA-P And SIMCA-P+
North America:
Umetrics Inc.
17 Kiel Ave.
Kinnelon NJ 07405
USA
Phone: +1 973 492 8355
Fax: +1 973 492 8359
Email: info.us@umetrics.com
IID 2029 2005

General points
• Data can be compiled in Excel or as comma separated values (.csv)
• Data may be held in Access database files
• If your data has replicates it is usually best to include these rather than averaging them.
• Process data should be collected at the same time points.
• Small amounts of missing data can be tolerated.
1. QSAR
Compound 1
Compound 2
Compound 3
Molecular or Physical Properties Activity data
Var1 Var2 Var3 Var4 Act 1 Act 2 Act 3
2. Genomics / Proteomics / Metabonomics
If you have more than 256 gene / protien / NMR columns and you are using Excel you will have to initially format the
table with the gene data arranged vertically. Data can then be transposed during SIMCA-P import. Otherwise use
Comma separated value files (.CSV).
Dose 1
Dose 2 Fluorescence data
Dose 3
Dose n
Gene / Protein / Peak Dose
Gene / Protein / Peak Dose
Drug 1 Dose 1
Drug 1 Fluorescence data Dose 2
Drug 2 Dose 1
Drug 2 Dose 2
Dose n Dose n
3. Spectroscopic data
Most spectroscopic instruments can save data in one of the following formats that SIMCA-P supports JCAMP-DX,
MVACDF, Brimrose files, Galactic SPC files, NSAS files. Saving them like this is your first choice.
For spectral data with less than 256 variables (wavelengths) they can also be saved in Excel as below.
Spectra 1
Spectra 2 Spectroscopy data
Spectra n
For spectra with more than 256 variables (wavelengths) format with spectral data going down in columns to overcome
the maximum column width of 256 in Excel. Data is then transposed during SIMCA-P import
UMETRICS GUIDING THE WAY – USING SIMCA-P 2
Y-var

Example: NIR spectra of raw material:
Spectrum
Batch 1
Spectrum
Batch 2
Spectrum
Batch n
Wavelength Absorbance Absorbance Absorbance
210 0.12 0.45 0.34
211 0.45 0.12 0.23
212….etc 0.23 0.23 0.12
Example: Comparing spectra of different compounds:
Compound 1 Compound 2 Compound n
Wavelength Absorbance Absorbance Absorbance
210 0.12 0.45 0.34
211 0.45 0.12 0.23
212….etc 0.23 0.23 0.12
Property or activity 1
Property or activity 2
4. Process data (or biological time courses)
Monitoring a process:
Time 1
Time 2
Time 3
Tag description
(ex. pressure pos2)
Tag name
(ex. FFC2-p002-xx01)
Variables
Equal distances between time samples is preferable if it is a continuous process. The identity of the measured variables
are preferable in two or more cells in each column, one for an interpretable name and one for the technical tag
name in the database.
Modelling a process output:
Time 1
Time 2
Time n
Input Measurements Quality or Output Measurements
For batch data (for metabonomics data each batch is an individual):
Time
Time
Variables
Batch 1
Batch 2
See the Appendix for a detailed description of how to format batch data.
UMETRICS GUIDING THE WAY – USING SIMCA-P 3

5. LC-MS or GC-MS data
LC or GC –MS data is in the form of a 3 dimensional data cube. The way to format this data is to unfold it so that
each variable becomes a time:mass paired variable. Software such as Marker Lynx from Waters and Metalign from
Plant Research International can be used for this.
Sample
Sample
Sample 1
Sample 2
Sample 3
Time
1.0min:245 m/z 1.0min:345 m/z 1.2min:567 m/z 3.4min:678 m/z
6. Chromatography & Electrophoresis data
m/z
T1 T1 T1 T1
HPLC, GC and CE data is best used as peak tables using the chromatographic instrument software to deal with small
shifts in retention times by means of retention time windowing. Even slight shifts in retention time will hinder chemometric
analysis severely if the raw data is used.
Peak1 Peak2 Peak3 Peak4
Sample1 1.0% 33.6% 25.5% 2.7%
Sample2 2.3% 23.1% 34.5% 10.8%
Sample3 5.4% 26.8% 41.6% 5.4%
2-D Electrophoresis Gels and Proteomics data is best analysed as peak tables constructed by dedicated image analysis
software before import into SIMCA-P.
UMETRICS GUIDING THE WAY – USING SIMCA-P 4

Detailed batch data description
APPENDIX
1. Compile initial conditions, batch evolution and final quality data.
2. Consider every single source of variation:
• What went in to the batch?
• What has been done to the batch?
This includes SOP logs, recipes, provenance of raw materials, process historian data, process logs, metadata
(operator, equipment, reactor, site) etc.
3. Compile the data in a format similar to the table below:
UMETRICS GUIDING THE WAY – USING SIMCA-P 5