Skip to content

Latest commit

 

History

History
61 lines (33 loc) · 3.07 KB

analyses.md

File metadata and controls

61 lines (33 loc) · 3.07 KB
Raw

Analyses

Analyses are automatically started according to the "Families to analyze" queue. The family is then given an action of "running".

@ rasta:~/servers/crontab/analysis-auto.sh
$ cg analysis auto
$ cg analysis -f FAMILY-ID

This process involves a set of steps that setup the environment to prepare for analysis start:

  • config: a pedigree config is created in a "family" directory in the analysis root folder. Information is mainly fetched from status with some exceptions like picking up capture kit version which is stored in LIMS.

    cg analysis config FAMILY-ID

  • link: link and rename FASTQ files for each sample in the family. We rely on all FASTQ files to already be stored in Housekeeper for each sample. Information to rename samples and link paried-end reads is fetched from the headers in the FASTQ-file itself.

    cg analysis link -f FAMILY-ID

  • panel: generate a gene panel with relevant clinical genes for the analysis. The latest version of the "OMIM-AUTO" panel is seeded with every analysis to avoid errors with very few variants in the output. Panels information is managed in Scout. For collaborating customers: cust002-4 we run a pre-defined set of panels for most cases.

    cg analysis panel FAMILY-ID

  • start: just start the pipeline with existing config/FASTQs/panel. This can be useful if you need to make custom edits to the config that you want to run with.

    cg analysis start FAMILY-ID

The analysis is thereafter tracked in Trailblazer until it's successfully completed. An analysis can fail and be restarted but when it eventually completes it is automatically picked up first by Trailblazer.

@ rasta:~/servers/crontab/trailblazer-scan.sh
$ trailblazer scan

Relevant files from all completed analyses are then stored in Housekeeper. An analysis record is also created in status in this process. Finally, the action ("running") for the related family is reset.

@ rasta:~/servers/crontab/store-completed.sh
$ cg store completed
$ cg store analysis [ANALYSIS-CONFIG-PATH]

The files to store are defined in the Trailblazer API in cg.

Analysis output of completed runs should be cleaned up after two weeks if it's been stored in Housekeeper.

cg clean auto

The command will promt you for confirmation before deleting any files.

Failed runs

analysisrunstatus

Failure in this step indicates a quality-related error. The next step is usually to make an assesment of the deviation. You can start by running:

trailblazer check FAMILY-ID

This command will report sex determination and duplicates for all samples.

Saliva samples

Missing FASTQ files

Poor coverage can be caused by missing FASTQ files that weren't linked before analysis start. It could also be a result of certain sample origins, e.g. "saliva" where it's expected to see a higher number of duplicates and mapping to bacterial genome sequences. You can find out the origin of the sample in the "Source" sample UDF in LIMS.