Merge branch 'devel' of https://github.com/gis-rpd/pipelines into devel

README.md

@@ -39,12 +39,12 @@ The following installations are available at different sites (referred to as `RP
 - NSCC: `/home/users/astar/gis/gisshared/rpd/pipelines/`
 
 Each of these contains one subfolder per pipeline version,
-e.g. `$RPD_PIPELINES/pipelines.2017-01` (referred to as
+e.g. `$RPD_PIPELINES/pipelines.2017-06` (referred to as
 `PIPELINE_ROOTDIR` below).
 
 Much of this framework assumes a certain setup and services to be
 present, as is the case in GIS / the NSCC. This repository is
-therefore of limited use to the general public. See INSTALL.md for
+therefore of limited use to the general public. See `INSTALL.md` for
 simplistic installation instructions.
 
 Some pipelines only work at a certain site (due to system or software
@@ -112,23 +112,25 @@ In either case, you must not prefix the script with `python`.
 
 | Name | Category | Notes | @GIS | @NSCC |
 | ---  | ---      | ---   | ---  | ---   |
-| [bcl2fastq](bcl2fastq/README.md)                                    | Production          | Not for end-users     | Y | Y |
-| [ChIP-seq](chromatin-profiling/chipseq/README.md)                   | Chromatin Profiling |                       | Y | Y |
-| [SG10K](custom/SG10K/README.md)                                     | Custom              | Not for end-users     | Y | Y |
-| [ViPR](germs/vipr/README.md)                                        | GERMS               |                       | Y | Y |
-| [BWA-MEM](mapping/BWA-MEM/README.md)                                | Mapping             |                       | Y | Y |
+| [bcl2fastq](bcl2fastq/README.md)            | Production          | Not for end-users     | Y | Y |
+| [ATAC-seq](chromatin-profiling/atacseq/README.md)             | Chromatin Profiling |                       | Y | Y |
+| [ChIP-seq](chromatin-profiling/chipseq/README.md)             | Chromatin Profiling |                       | Y | Y |
+| [SG10K](custom/SG10K/README.md)                | Custom              | Not for end-users     | Y | Y |
+| [ViPR](germs/vipr/README.md)                 | GERMS               |                       | Y | Y |
+| [BWA-MEM](mapping/BWA-MEM/README.md)              | Mapping             |                       | Y | Y |
 | [Shotgun Metagenomics](metagenomics/shotgun-metagenomics/README.md) | Metagenomics        |                       | Y | Y |
-| [Essential-Genes](metagenomics/essential-genes/README.md)           | Metagenomics        | Requires ref download | Y | Y |
-| [STAR-RSEM](rnaseq/star-rsem/README.md)                             | RNA-Seq             |                       | Y | Y |
-| [Fluidigm-HT-C1-RNASeq](rnaseq/fluidigm-ht-c1-rnaseq/README.md)     | RNA-Seq             |                       | Y | N |
-| [LoFreq-Somatic](somatic/lofreq-somatic/README.md)                  | Somatic             |                       | Y | N |
-| [Mutect](somatic/mutect/README.md)                                  | Somatic             |                       | Y | Y |
-| [GATK](variant-calling/gatk/README.md)                              | Variant-calling     |                       | Y | Y |
-| [Lacer-LoFreq](variant-calling/lacer-lofreq/README.md)              | Variant-calling     |                       | Y | N |
+| [Essential-Genes](metagenomics/essential-genes/README.md)      | Metagenomics        | Requires ref download | Y | Y |
+| [STAR-RSEM](rnaseq/star-rsem/README.md)            | RNA-Seq             |                       | Y | Y |
+| [Fluidigm-HT-C1-RNASeq](rnaseq/fluidigm-ht-c1-rnaseq/README.md)| RNA-Seq             |                       | Y | N |
+| [Wafergen](rnaseq/wafergen/README.md)             | RNA-Seq             | Requires cellular barcodes | Y | Y |
+| [LoFreq-Somatic](somatic/lofreq-somatic/README.md)       | Somatic             |                            | Y | N |
+| [Mutect](somatic/mutect/README.md)               | Somatic             |                            | Y | Y |
+| [GATK](variant-calling/gatk/README.md)                 | Variant-calling     |                            | Y | Y |
+| [Lacer-LoFreq](variant-calling/lacer-lofreq/README.md)         | Variant-calling     |                            | Y | N |
 
 See `example-dag.pdf` in each pipeline's folder for a visual overview of the workflow.
 
-Note, pipelines start with fastq files as input (a few allow injection of BAM files).
+Note, most pipelines start with FastQ files as input, a few allow injection of BAM files.
 
 ## How it Works
 
@@ -138,7 +140,7 @@ Note, pipelines start with fastq files as input (a few allow injection of BAM fi
   `conf.yaml` file) and gets its own readgroup assigned where
   appropriate.
 - Software versions are defined in each pipelines' `cfg/modules.yaml`
-  and loaded via [dotkit](https://computing.llnl.gov/?set=jobs&page=dotkit)
+  and loaded via [Lmod](http://lmod.readthedocs.io/en/latest/)
 - Pipeline wrappers create an output directory containing all
   necessary configuration files, run scripts etc.
 - After creation of this folder, the analysis run is automatically submitted to the cluster
@@ -154,7 +156,6 @@ Note, pipelines start with fastq files as input (a few allow injection of BAM fi
 
 First call the wrapper in question with `--no-run`. cd into the given outdir and then
 - Check the created `conf.yaml`
-- Print the DAG: `rm -f logs/snakemake.log; type=pdf; EXTRA_SNAKEMAKE_ARGS="--dag" bash run.sh; cat logs/snakemake.log | dot -T$type > dag.$type`
 - Execute a dryrun: `rm -f logs/snakemake.log; EXTRA_SNAKEMAKE_ARGS="--dryrun" bash run.sh; cat logs/snakemake.log`
 - Run locally: `nohup bash run.sh; tail -f logs/snakemake.log`
 
@@ -189,11 +190,11 @@ described in the following:
   want to set the CSV delimiter with `-d`, e.g. `-d ,`
 - Use the created yaml file as input for the pipeline wrapper (option `--sample-cfg your.yaml`)
 
-Please note, not all pipelines support this feature (for example the
-somatic pipelines don't), but most do, e.g. GATK, Lacer-LoFreq. In
-some cases multisample processing can lead to very high memory
-consumption by the snakemake master process itself, a side-effect
-which is hard to predict.
+Please note, not all pipelines support this feature, for example the
+Chipseq and all somatic pipelines. In some cases multisample
+processing can lead to very high memory consumption by the snakemake
+master process itself, a side-effect which is hard to predict (the
+master process will be killed).
 
 The above configuration can be used for single sample processing as
 well, however, for single samples the corresponding use of options

VERSION

@@ -1 +1 @@
-2017-06.0
+2017-10.0

bcl2fastq/Snakefile

@@ -23,6 +23,7 @@ from pipelines import send_status_mail
 from pipelines import path_to_url
 from pipelines import RPD_SIGNATURE
 from utils import generate_timestamp
+from pipelines import mark_as_completed
 from elmlogger import ElmLogging, ElmUnit
 from bcl2fastq_dbupdate import DBUPDATE_TRIGGER_FILE_FMT, DBUPDATE_TRIGGER_FILE_MAXNUM
 from readunits import sampledir_to_cfg
@@ -149,6 +150,8 @@ onsuccess:
                          os.path.abspath(RESULT_OUTDIR), extra_text=extra_text)
     # cannot talk to mongodb from compute. use trigger file
     write_db_update_trigger(True)
+
+    mark_as_completed()
 
 
 onerror:

changelog.md

@@ -3,6 +3,26 @@
 This change log only lists the major changes between releases. For a
 full list of changes refer to the commit log.
 
+## 2017-10
+
+New pipelines:
+- chromatin-profiling/atacseq: runs Bowtie2 and MACS2 to call ATAC-Seq peaks
+- rnaseq/wafergen-scrna: analyses WaferGen's single cell sequencing
+  data, using umis and scRNApipe for the core analysis.
+
+
+Changes to pipelines and framework:
+- Major pipelines now produce benchmarking logs
+- STAR-RSEM now supports --estimate-rspd and new strandedness option
+- Shotgun-metagenomics: Added SRST2, read counting and coverage
+  threshold for decontamination
+- GATK: support for joint variant calling (--joint-calls) and added a
+  new option (--gvcf-only). Also added a new optional references
+  config for agressive splitting
+- Added --restarts option to control number of automatic restarts in
+  case of failure
+- Minor additions to SG10K and many under the hood changes to bc2lfastq
+
 ## 2017-06
 
 New pipelines:

chromatin-profiling/atacseq/tests.sh

@@ -68,7 +68,7 @@ echo 'FIXME test quality of results' 1>&2
 
 
 # DAG
-SKIP_DAG=0
+SKIP_DAG=1
 if [ $SKIP_DAG -eq 0 ]; then
     echo "DAG: PE two pairs" | tee -a $log
     odir=$($DOWNSTREAM_OUTDIR_PY -r $(whoami) -p $PIPELINE)

chromatin-profiling/chipseq/tests.sh

@@ -78,7 +78,7 @@ echo 'FIXME bam injection' 1>&2
 
 
 # DAG
-SKIP_DAG=0
+SKIP_DAG=1
 if [ $SKIP_DAG -eq 0 ]; then
     echo "DAG: WES" | tee -a $log
     odir=$($DOWNSTREAM_OUTDIR_PY -r $(whoami) -p $PIPELINE)

custom/SG10K/Snakefile

@@ -134,9 +134,11 @@ rule bam_to_cram:
     #conda:
     #    "env.yaml"
     shell:
-        "module load bamutil/1.0.14-33-nonprimdup;"
+        "{{ "
+        " module load bamutil/1.0.14-33-nonprimdup;"
         " bam squeeze --in {input.bam} --out -.ubam {params.bin_arg} |"
-        "samtools view -C -T {input.reffa} -@ {threads} -o {output.cram} - >& {log}"        
+        " samtools view -C -T {input.reffa} -@ {threads} -o {output.cram} -;"
+        " }} >& {log}"        
 
 
 localrules: cram_index        

custom/SG10K/cfg/modules.yaml

@@ -7,4 +7,3 @@ datamash: 1.1.0
 sg10k-cov: '062017'
 fastmitocalc: 'default'
 
-

custom/SG10K/tests.sh

@@ -61,7 +61,7 @@ echo "Check log if the following final message is not printed: \"$COMPLETE_MSG\"
 
 
 # DAG
-SKIP_DAG=0
+SKIP_DAG=1
 if [ $SKIP_DAG -eq 0 ]; then
     echo "DAG: $SAMPLE" | tee -a $log
     odir=$($DOWNSTREAM_OUTDIR_PY -r $(whoami) -p $PIPELINE)

germs/vipr/aux/plot.py

@@ -36,7 +36,7 @@ def parse_genomecov(genomecov_gzfile):
         for line in fh:
             sq, pos, cov = line.decode().rstrip().split("\t")
             pos = int(pos)-1
-            cov = int(cov)
+            cov = int(float(cov))# float for support of scientific notation
             if sq not in genomecov:
                 genomecov[sq] = OrderedDict()
             genomecov[sq][pos] = cov

germs/vipr/tests.sh

@@ -69,7 +69,7 @@ base_cmd="$WRAPPER -1 $FQ1 -2 $FQ2 -s test:DENV2-TSV01-PDH203 -r $REF";
 # FIXME fix here and elsewhere
 
 # DAG
-SKIP_DAG=0
+SKIP_DAG=1
 if [ $SKIP_DAG -eq 0 ]; then
     echo "Creating DAG" | tee -a $log
     odir=$($DOWNSTREAM_OUTDIR_PY -r $(whoami) -p $PIPELINE)

mapping/BWA-MEM/tests.sh

@@ -75,7 +75,7 @@ echo "Check log if the following final message is not printed: \"$COMPLETE_MSG\"
 
 
 # DAG
-SKIP_DAG=0
+SKIP_DAG=1
 if [ $SKIP_DAG -eq 0 ]; then
     echo "DAG: PE through config" | tee -a $log
     odir=$($DOWNSTREAM_OUTDIR_PY -r $(whoami) -p $PIPELINE)

metagenomics/essential-genes/tests.sh

@@ -73,7 +73,7 @@ echo "Check log if the following final message is not printed: \"$COMPLETE_MSG\"
 
 
 # DAG
-SKIP_DAG=0
+SKIP_DAG=1
 if [ $SKIP_DAG -eq 0 ]; then
     echo "DAG" | tee -a $log
     odir=$($DOWNSTREAM_OUTDIR_PY -r $(whoami) -p $PIPELINE)

metagenomics/shotgun-metagenomics/tests.sh

@@ -64,7 +64,7 @@ base_cmd="$WRAPPER  --sample-cfg $SAMPLECFG --name test:shotgun-metagenomics";
 # FIXME fix here and elsewhere
 
 # DAG
-SKIP_DAG=0
+SKIP_DAG=1
 if [ $SKIP_DAG -eq 0 ]; then
     echo "Creating DAG" | tee -a $log
     odir=$($DOWNSTREAM_OUTDIR_PY -r $(whoami) -p $PIPELINE)

rnaseq/fluidigm-ht-c1-rnaseq/tests.sh

@@ -65,7 +65,7 @@ cmd_base="$WRAPPER -1 $COL1_R1 -2 $COL1_R2 -s COL01 --name 'test:COL01'"
 
 
 # DAG
-SKIP_DAG=0
+SKIP_DAG=1
 if [ $SKIP_DAG -eq 0 ]; then
     echo "DAG" | tee -a $log
     odir=$($DOWNSTREAM_OUTDIR_PY -r $(whoami) -p $PIPELINE)

rnaseq/star-rsem/tests.sh

@@ -84,7 +84,7 @@ CMD_FULL="$WRAPPER -1 $R1_FULL -2 $R2_FULL -s $SAMPLE --name 'test:FULL'"
 SKIP_REAL_FULL=0
 
 # DAG
-SKIP_DAG=0
+SKIP_DAG=1
 if [ $SKIP_DAG -eq 0 ]; then
     echo "DAG: Full" | tee -a $log
     odir=$($DOWNSTREAM_OUTDIR_PY -r $(whoami) -p $PIPELINE)

rnaseq/wafergen-scrna/README.md

@@ -54,8 +54,7 @@ Reports:
 
 ## Notes
 
-Deduplication can be very slow for large data-sets. If deduplication
- is not a must, we recommend to switch it off (`--no-dedup`).
+Deduplication can be very slow for large data-sets. We recommend to not use it unless necessary (`--dedup`).
 
 ## References
 

rnaseq/wafergen-scrna/Snakefile

@@ -192,6 +192,8 @@ rule bamtag_split:
     output:
         taggedbam = '{prefix}/{sample}/bamtag/{sample}_R2.tagged.bam',
         splitflag = touch('{prefix}/{sample}/bamtag/{sample}_R2.tagsplit.COMPLETE')
+    log:
+        '{prefix}/{sample}/bamtag/{sample}_R2.tagged.log'
     benchmark:
          '{prefix}/{sample}/bamtag/{sample}_R2.tag.bamtag_split.benchmark.log'
     params:
@@ -201,10 +203,12 @@ rule bamtag_split:
     threads:
         1
     shell:
-        'umis bamtag {input.bam} | samtools addreplacerg'
+        '{{'
+        ' umis bamtag {input.bam} | samtools addreplacerg'
         ' -r ID:{params.sample} -r LB:{params.sample} -r SM:{params.sample} -r PL:{params.platform} -r PU:1 -r CN:{params.center}'
         ' -o - - | samtools sort -o {output.taggedbam} -T {output.taggedbam}.tmp -;'
-        ' bamtools split -tag XC -in {output.taggedbam}'
+        ' bamtools split -tag XC -in {output.taggedbam};'
+        ' }} >& {log}'
         # not guaranteed to create one file per barcode
 
 

rnaseq/wafergen-scrna/tests.sh

@@ -63,7 +63,7 @@ cmd_base="$WRAPPER -c $TEST_CBINDEX -S $TEST_SAMPLECFG"
 
 
 # DAG
-SKIP_DAG=0
+SKIP_DAG=1
 if [ $SKIP_DAG -eq 0 ]; then
     echo "DAG" | tee -a $log
     odir=$($DOWNSTREAM_OUTDIR_PY -r $(whoami) -p $PIPELINE)
@@ -103,9 +103,9 @@ if [ $skip_real_runs -ne 1 ]; then
     jid=$(tail -n 1 $odir/logs/submission.log  | cut -f 3 -d ' ')
     echo "Started job $jid writing to $odir. You will receive an email"
 
-    echo "Realrun no_dedup" | tee -a $log
+    echo "Realrun dedup" | tee -a $log
     odir=$($DOWNSTREAM_OUTDIR_PY -r $(whoami) -p $PIPELINE)
-    eval $cmd_base --no-dedup -o $odir -v >> $log 2>&1
+    eval $cmd_base --dedup -o $odir -v >> $log 2>&1
     # magically works even if line just contains id as in the case of pbspro
     jid=$(tail -n 1 $odir/logs/submission.log  | cut -f 3 -d ' ')
     echo "Started job $jid writing to $odir. You will receive an email"

rnaseq/wafergen-scrna/wafergen-scrna.py

@@ -78,8 +78,8 @@ def main():
     d = 20.0
     parser.add_argument('--frag-len-sd', default=d, type=float,
                         help="Estimated fragment length standard deviation (default={})".format(d))
-    parser.add_argument('--no-dedup', action="store_true",
-                        help="Skip UMI-based deduplication (can be slow)")
+    parser.add_argument('--dedup', action="store_true",
+                        help="Run UMI-based deduplication (slow for large data-sets!)")
     args = parser.parse_args()
 
     # Repeateable -v and -q for setting logging level.
@@ -139,7 +139,7 @@ def main():
     cfg_dict['cell_barcodes'] = os.path.abspath(args.cell_barcodes)
     cfg_dict['frag_len'] = args.frag_len
     cfg_dict['frag_len_sd'] = args.frag_len_sd
-    cfg_dict['no_dedup'] = args.no_dedup
+    cfg_dict['no_dedup'] = not args.dedup
     cfg_dict['scrnapipe_transform'] = os.path.abspath(os.path.join(
         PIPELINE_BASEDIR, 'aux/transform.json'))
     cfg_dict['scrna_conf_template'] = os.path.abspath(os.path.join(

somatic/lofreq-somatic/tests.sh

@@ -76,7 +76,7 @@ wgs_cmd_base="$WRAPPER --normal-bam $DREAM_WGS_NORMAL_BAM --tumor-bam $DREAM_WGS
 
 
 # DAG
-SKIP_DAG=0
+SKIP_DAG=1
 if [ $SKIP_DAG -eq 0 ]; then
     echo "DAG: WES" | tee -a $log
     odir=$($DOWNSTREAM_OUTDIR_PY -r $(whoami) -p $PIPELINE)

somatic/mutect/tests.sh

@@ -78,7 +78,7 @@ wgs_cmd_base="$WRAPPER --normal-bam $DREAM_WGS_NORMAL_BAM --tumor-bam $DREAM_WGS
 
 
 # DAG
-SKIP_DAG=0
+SKIP_DAG=1
 if [ $SKIP_DAG -eq 0 ]; then
     echo "DAG: WES" | tee -a $log
     odir=$($DOWNSTREAM_OUTDIR_PY -r $(whoami) -p $PIPELINE)

tools/pipelint.py

@@ -111,6 +111,7 @@ def check_modules(pipeline_dir):
 
     is_ok = True
     module_cfgs = glob.glob(os.path.join(pipeline_dir, "cfg/modules.yaml"))
+    assert len(module_cfgs) > 0
     modules = dict()
     for cfg in module_cfgs:
         with open(cfg) as fh:
@@ -139,7 +140,14 @@ def main(pipelinedirs,
     logger.warning("include other existing tools here: check_cluster_conf.py...")
 
     snakefiles = [os.path.join(d, "Snakefile") for d in pipelinedirs]
-
+
+    if not no_modules_check:
+        for d in pipelinedirs:
+            if not check_modules(d):
+                print("FAILED: Modules check for {}".format(d))
+            else:
+                print("OK: Modules check for {}".format(d))
+
     includes = []
     for f in snakefiles:
         assert os.path.exists(f)
@@ -148,12 +156,6 @@ def main(pipelinedirs,
         else:
             print("OK: Expected files for {}".format(f))
 
-        if not no_modules_check:
-            if not check_modules(f):
-                print("FAILED: Modules check for {}".format(f))
-            else:
-                print("OK: Modules check for {}".format(f))
-
         includes.extend(get_includes_from_snakefile(f))
 
 

variant-calling/gatk/tests.sh

@@ -80,7 +80,7 @@ wgs_cmd_base="$WRAPPER -1 $WGS_FQ1 -2 $WGS_FQ2 -s NA12878-WGS -t WGS --name 'tes
 
 
 # DAG
-SKIP_DAG=0
+SKIP_DAG=1
 if [ $SKIP_DAG -eq 0 ]; then
     echo "DAG: WES" | tee -a $log
     odir=$($DOWNSTREAM_OUTDIR_PY -r $(whoami) -p $PIPELINE)

variant-calling/lacer-lofreq/tests.sh

@@ -76,7 +76,7 @@ wgs_cmd_base="$WRAPPER -1 $WGS_FQ1 -2 $WGS_FQ2 -s NA12878-WGS -t WGS --name 'tes
 
 
 # DAG
-SKIP_DAG=0
+SKIP_DAG=1
 if [ $SKIP_DAG -eq 0 ]; then
     echo "DAG: WES" | tee -a $log
     odir=$($DOWNSTREAM_OUTDIR_PY -r $(whoami) -p $PIPELINE)