Notes meeting 27-10-2010
Ideas
- Have a look at Lisa/Huygens? cluster/grid
- LISA -> M. Kattenberg will have a look
- Phillips cluster could be an option
- Leiden has shark cluster, 108 cpu's
- If all trios put together for SNP calling etc. have a look for repeated error patterns in Illumina machines
- Have a look at indel positions after calling, is position correct?
Action points
- Put BGI lab protocol on wiki, insertsizes etc
- Do check for interference etc
- Solve I/O issues by next week
- Add Pindel to pipeline
Questions
- Is QC recalibration also needed for other platforms?
Pipeline Utrecht
- Raw data
- Split by barcodes/samples
- Chunking and coversion to fastq
- Fastq sent to nodes
- Nodes return sorted .BAMs and statistics
- Main node merges all data from other nodes
- Custom pileup script merges (Frans-Paul)
- Annotation
- Query ENSEMBL from cluster nodes
- Main node merges all output, filter on region of interest, put result in database
- SNP calling
- GATK in progress
- VarScan? (Lot of false positives, not good enough)
Grid
- Two options:
- Commandline
- Workflow
Download in other formats: