From LipidomicsWiki
Contents |
SPP: Filtering of Affymetrix GeneChip-data
This SPP, describes the processing and filtering of Affymetrix GeneChip-data as delivered by the Software "Microarray Suite 5.0". Further SPPs will be added when the final microarray-plattform is confirmed.
Input:
Output of "Microarray Suite 5.0".
Output:
A annotated prefiltered list with microarray statistics.
Processing:
Addition of annotation
Download and match the Annotation-File with the Data-File to match the data with ENSEMBL-Code, Description, Symbol, Chromosomal Locus, etc.
Conversion of detection p-values into levels of detection
Dpv < 0.005 --> ++
0.005 <= Dpv < 0.05 --> +
0.05 <= Dpv --> -
Conversion of Signal log ratios into Folds
Fold = 2Signal log ratio
Addition of summary statistics:
Addtion of the following columns
- number of + (count the number of plus signs in every line)
- all absent (Is Detection always “A”=“absent”?)
- all NC (Is Change always “NC” = “No Change”?)
- sum of signals
- 2fold (To choose the differentially expressed genes look at every pair of Fold and Change in the list. If there exist at least one pair with Change ≠ “NC” and Fold ≥ 1.95 (if rounded to the first digit behind the dot =2.0) then write “2fold”, else write “del”)
Actual version of the gene-list:
Deletion of nonessential columns:
Deletion of the following columns:
*Detection, *Detection p-value, *Signal log ratio, *Change p-value
Find “Best ENSEMBL ProbeSet” (“BEP”)
Sometimes more than one probe-set exists for a certain gene. The one with the highest “number of +” and the highest “sum of signal” is chosen.
Final version of the gene-data:
Match the “BEP”-List to whatever Genelist you like
Deletion of rows with no 2fold genes
Delete all rows marked as "del".
Proposals and discussion points
Write here any proposals and discussion points you may have.







