source: extensions/net.sf.basedb.reggie/trunk/reggie-ogs-hosts.xml @ 2553

Last change on this file since 2553 was 2553, checked in by Nicklas Nordborg, 7 years ago

References #614: Improve error handling when executing jobs on the cluster

Started to incorporate alignment step into main script. The call to to tophat seems to be working, but there are several additional steps from the tophat_single.sh script that remains.

File size: 4.4 KB
Line 
1<?xml version="1.0" encoding="UTF-8"?>
2<reggie>
3
4<ogs-hosts>
5  <!-- one or more hosts entries. Each entry define a Open Grid Scheduler cluster -->
6  <!-- Add entries as needed and store in <base-dir>/www/WEB-INF/classes/reggie-ogs-hosts.xml -->
7
8  <host 
9    name="Display name of host"
10    address="ip-name.of.the.host"
11    port="22"
12    public-key="Base64-encoded public key of SSH server"
13    user=""
14    password=""
15    >
16    <!-- job-folder: must be a path that is common to the master and all nodes (required) -->
17    <job-folder>/path/to/job/folder</job-folder>
18    <!-- tmp-folder: can be a local path on the node but must exist on all nodes (required) -->
19    <tmp-folder>/tmp</tmp-folder>
20    <!-- script-folder: folder where the pipeline scripts are located (required).
21      Must be accessible from all nodes.
22     -->
23    <script-folder>/path/to/script/folder</script-folder>
24   
25    <!-- Full path to the location where data files should be archived (required) -->
26    <!-- Do not include name of project, since it is added by the Reggie wizard -->
27    <project-archive>/path/to/project-archive</project-archive>
28   
29    <!-- Full path to the root location where reference genomes are located -->
30    <!-- Do not include name of project -->
31    <reference-folder>/disk/reference</reference-folder>
32   
33    <!-- priority values that are selectable in the web interface -->
34    <!-- allowed range is -1023 to 1024 -->
35    <!-- NOTE! positive values require special permissions on the cluster -->
36    <priorities>
37      <!-- <priority name="high" value="500" /> -->
38      <priority name="normal" value="0" default="true" />
39      <priority name="low" value="-500" />
40    </priorities>
41   
42    <!-- settings for the demuxing step -->
43    <demux>
44      <!-- full path to the directory with Picard jar files (required) -->
45      <picard-dir></picard-dir>
46      <!-- amount of memory to give to Picard (default is 50g)-->
47      <picard-memory>50g</picard-memory>
48      <!-- parallel environment option to the queue system -->
49      <!-- the default setting use all slots on hosts with at least 8 slots available -->
50      <parallel-environment>smp 8-</parallel-environment>
51      <!-- full path to the location where HiSeq data is stored (required) -->
52      <run-archive>/path/to/run_archive</run-archive>
53      <!-- static options for the picard ExtractIlluminaBarcodes step -->
54      <extract-options>QUIET=true VERBOSITY=WARNING</extract-options>
55      <!-- static options for the picard IlluminaBasecallsToFastq step -->
56      <fastq-options>ADAPTERS_TO_CHECK=null ADAPTERS_TO_CHECK=INDEXED ADAPTERS_TO_CHECK=PAIRED_END INCLUDE_NON_PF_READS=false MAX_READS_IN_RAM_PER_TILE=5000000 QUIET=true VERBOSITY=WARNING</fastq-options>
57      <!-- number of tiles to process when debugging (default=16) -->
58      <debug-tile-limit>16</debug-tile-limit>
59    </demux>
60
61    <!-- Settings for the Trimmomatic step -->
62    <trimmomatic>
63      <!-- full path to the JAR file with the Trimmomatic program -->
64      <jar-path>/usr/local/packages/trimmomatic/0.32/trimmomatic-0.32.jar</jar-path>
65      <!-- full path to the file with Illumina adapter information -->
66      <adapter-file>/usr/local/packages/trimmomatic/0.32/adapters/TruSeq3-PE-2.fa</adapter-file>
67      <!-- static options for Trimmomatic -->
68      <options>ILLUMINACLIP:${AdapterFile}:2:30:12:1:true LEADING:3 TRAILING:3 MAXINFO:40:0.9 MINLEN:20</options>
69    </trimmomatic>
70
71    <pe-filter>
72      <!-- static options for bowtie -->
73      <bowtie-options>-q --fr -k 1 --phred33 -t --local</bowtie-options>
74
75      <!-- parallel environment option to the queue system -->
76      <!-- the default setting use all slots on hosts with at least 8 slots available -->
77      <parallel-environment>smp 8-</parallel-environment>
78
79      <!-- relative path from <reference-folder> to the reference genome used for filtering -->
80      <!-- TODO selectable in GUI? saved as annotation? -->
81      <reference-name>scanb/ribo_phix_repeats_filter.bt2/ribo_phix_repeats_filter</reference-name>
82    </pe-filter>
83
84    <tophat>
85      <!-- relative path from <reference-folder> to the reference genome used for alignment -->
86      <!-- TODO selectable in GUI? saved as annotation? -->
87      <reference-gidx>scanb/hg19_1-22_XM_extra_b37_Y_decoy.bt2/hg19_1-22_XM_extra_b37_Y_decoy</reference-gidx>
88      <reference-tidx>scanb/UCSC_Human_hg19_knownGenes_GTF_appended_10sep2012v209.bt2/knownGenes</reference-tidx>
89      <!-- static options for tophat -->
90      <options>--library-type fr-firststrand --keep-fasta-order --no-coverage-search</options>
91    </tophat>
92  </host>
93
94 
95</ogs-hosts>
96
97</reggie>
Note: See TracBrowser for help on using the repository browser.