Skip to content

Latest commit

 

History

History
212 lines (130 loc) · 6.15 KB

CHANGELOG.md

File metadata and controls

212 lines (130 loc) · 6.15 KB

Change Log

0.2.4 - 2016-07-06

Fixed

  • Test that option -k works as it should.

0.2.3 - 2016-06-06

Fixed

  • Delete the intermediate collated files only after they are successfully collated into one master file.

  • Prefer picardmetrics.conf in the working directory if it exists.

Changed

  • Increase compatibility:

    • Sample name is now the filename without the file extension.

    • Use find to discover Picard output files in directories.

    • Use the BAM dirname as part of the TEMP_DIR.

    • Call the output BAM file.MarkDuplicates.bam

    • Search for "rRNA" instead of gene_type "rRNA" to find ribosomal intervals in a GTF file.

0.2.2 - 2015-06-12

Fixed

  • Set TMP_DIR option in all invocations to Picard tools.

0.2.1 - 2015-06-05

Added

  • Add PREFIX-insert-size-histogram.tsv to the collated outputs.

  • Add PREFIX-base-distribution-by-cycle-histogram.tsv to the collated outputs.

Changed

  • Use ## METRICS and ## HISTOGRAM as landmarks to find the correct lines in the Picard output files. This is more robust than using line numbers.

  • Delete the collated output files that are successfully integrated into the PREFIX-all-metrics.tsv file. They are redundant.

0.2.0 - 2015-05-28

Added

  • Added two commands:

    • refFlat
    -   This creates a file for the REF_FLAT argument of
    CollectRnaSeqMetrics.
    • rRNA
    -   This creates a file for the RIBOSOMAL_INTERVALS argument of
    CollectRnaSeqMetrics.

  • Travis continuous integration testing. Each time I push new changes to the repo, Travis will automatically check that the picardmetrics output is correct.

Fixed

  • This was harmless, but some of the collated output files were missing the trailing tab character at the end of the line. Some also had extra lines without data. Now all of the collated output files should be perfectly formatted.

  • Bash functions return 1 when problems are encountered.

  • The -k option is checked.

Changed

  • picardmetrics run now puts the newly generated BAM file in /tmp and deletes it after successfully running, unless the -k option is used.

  • picardmetrics run now runs CreateSequenceDictionary and ReorderSam on the BAM file to avoid issues where the sequence names in the header of the BAM file is not in the same order as the sequence names in the reference sequence FASTA file.

  • Shuffled the header for data/project1/sample2.bam so we can ensure that BAM files with unordered headers work correctly.

  • Rearranged output files so you can see the extra metrics contributed by CollectRnaSeqMetrics:

    • Compare the files in out/default with out/rnaseq.

Removed

  • Removed files that are no longer needed:

    • scripts/make_refFlat
    • scripts/make_rRNA_intervals
    • scripts/plot_picardmetrics.R

0.1.4 - 2015-05-22

Fixed

  • The picard collate command now tolerates missing files and will still collate the existing files.

  • The collated output file PREFIX-alignment-metrics.tsv no longer has extra blanks lines.

  • Fix the invocation of md5sum -c in test/test.sh.

  • Some systems like CentOS don't have an up-to-date version of coreutils. So, I replaced the invocation of realpath with a Bash function that does produces the same output.

Changed

  • The rnaseq_metrics function will generate a sequence dictionary for the rRNA intervals_list file on-the-fly by taking it from the BAM file. The sequence dictionary consists of the BAM header lines starting with '@SQ'.

    • The script scripts/make_rRNA_intervals no longer downloads chromosome sizes from UCSC.

  • The sort_sam function checks if the output BAM is correct.

  • The reference genome sequence is no longer copied to RAM.

  • The script scripts/make_refFlat no longer downloads the reference genome. The user must use their own anyway.

  • In scripts/install_deps, roll back to Picard 1.126 because of a conflicting change introduced in Picard 1.130. See here for more details: broadinstitute/picard#212

0.1.3 - 2015-04-23

Fixed

  • The collate_ functions did not correctly check for the number of metrics files. This affects output generated using picardmetrics without -r.

Changed

  • The output BAM file is now deleted after a successful run.

  • Added option -k to optionally keep the output BAM file.

  • The configuration file is now called picardmetrics.conf instead of .picardmetricsrc.

Added

  • Added extra commands to the Makefile to ease installation of dependencies and download of human reference data.

    • Now h38 is downloaded by default instead of hg19. The user can modify the script to get different versions.

0.1.2 - 2015-04-09

Fixed

  • In collate_all, check for prefix-rnaseq-metrics.tsv file before attempting to merge the files into prefix-all-metrics.tsv.

    • Warnings are no longer printed when a user tries to collate output files after running picardmetrics without the '-r' option.

  • The tests have been updated: the md5sum of final output files is checked. This is more robust than checking content of log files.

Changed

  • Removed regular expression from is_bam function, so the script will run with Bash version 2 released prior to 2004.

  • Previously, output files were written to a directory called 'picardmetrics' created in the BAM file's directory. Currently, the output files are written to a user-specified directory or the current directory by default.

  • The collate_ functions previously accepted the BAM files output by picardmetrics as an argument. Now, the collate_ functions accept a directory instead and search for the appropriately named files. This means that the BAM files created by Picard may be deleted after running picardmetrics.

0.1.1 - 2015-03-20

Changed

  • Write output files to project1/sample1/picardmetrics instead of project1/sample1/.

0.1 - 2015-03-16

  • Initial release.