ArrayMap (AM) Processing SNP arrays of 9 different platforms

Summary of highlights

• with R packages aroma.affymetrix, DNAcopy
• start from .CEL files
• output total copy number (CN) probe files, segment files and allele-specific CN files and B allele frequency (BAF) segment files as well as LOH calculation and its segment files.
• for single samples, plot the genome-wide CN landscape, and also zooming in a specific region
• for multiple samples, plot aggregated CN segments by frequency (strip plot) and heatmap of CN landscape sorted by hierarchical clustering.

Examples utility scripts:

perl arrayplotter.pl -in test_file/GSM325151
perl arrayplotter.pl -in test_file/GSM412388
perl multiple_segment_plot.pl -f  test_file/multiple_segment_test.tsv -sf test_file/sample_types.tsv -genome hg19

##################################

Whole SNP array pipeline

##################################

Includes

• extract probe values from .CEL files
• segment with DNAcopy
• cleanup segment, adjust baseline
• make plots
• extract metadata from GEOmeta text files
• insert metadata to db
• insert probe, segment data to db
• update db if segments are re-processed
• update dbstats (not tested)

Directory structure

An example directory structure indicating one series with 2 array experiments, where array1 is processed with steps probe, segment and reseg, with results written into the processed directory.

working_dir/
├── PlatformInfo
├── ReferenceFile
├── annotationData
│   └── chipTypes
│       └── CytoScanHD_Array
│           └── CytoScanHD_Array.cdf
├── plmData
├── probeData
├── processed
│   └── series1
│       ├── array1
│       │   ├── probes,cn.tsv
│       │   ├── segments,cn,provenance.tsv
│       │   └── segments,cn.tsv
│       └── array2
└── rawData
    └── series1
        └── CytoScanHD_array
            ├── array1.CEL
            └── array2.CEL

Example usage

1. To segment all the (total copy number) probe files in one series, i.e. all files named probes,cn.tsv in working_dir/processed/series1/array.../ will be processed. The output files segments,cn.tsv will be written into the same array folder.

rscript --vanilla processPipeline_combined.r -w working_dir -s series1 -e segment -p cn

2. To do noise filtering on all the (total copy number) segment files in one series, i.e. all files named segments,cn.tsv in working_dir/processed/series1/array.../ will be processed. The original segments,cn.tsv will be renamed to segments,cn,provenance.tsv and the new segment files will be named as segments,cn.tsv and written into the same array folder.

rscript --vanilla processPipeline_combined.r -w working_dir -s series1 -e reseg -p cn

A particular array can be selected for this additional noise filtering process by -a

rscript --vanilla processPipeline_combined.r -w working_dir -s series1 -a array1 -e reseg -p cn

Notes

1. The metadata directory is /Volumes/arraymapIncoming/GEOmeta/
2. Default workingdir is ~/aroma/hg19/, for any step after probe, workingdir can be anything as long as samples in the following structure: $workingdir/processed/$series/$array. But probe step requires several other subdirectories in the workingdir: rawData, annotationData, PlatformInfo, referenceFile.
3. Default raw data retrieval directory is $workingdir/rawData in the structure series/array/xxx.CEL.