/usr/share/doc/coop-computing-tools/manual/sand.html is in coop-computing-tools-doc 4.0-2.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 | <html>
<head>
<title>SAND User's Manual</title>
</head>
<body>
<style type="text/css">
pre {
background: #ffffcc;
font-family: monospace;
font-size: 75%
font-align: left;
white-space: pre;
border: solid 1px black;
padding: 5px;
margin: 20px;
}
</style>
<h1>SAND User's Manual</h1>
<b>Last Updated January 2011</b>
<p>
SAND is Copyright (C) 2010- The University of Notre Dame.
This software is distributed under the GNU General Public License.
See the file COPYING for details.
<p>
<h2>Overview</h2>
<p>SAND is a set of modules for accelerating genome assembly and other
bioinformatics tasks.
Using the <a href="http://www.cse.nd.edu/~ccl/software/workqueue/">Work Queue</a>
framework, SAND can distribute computational tasks to hundreds of nodes harnessed
from clusters, clouds, grids, or just the idle machines you have in your office.
SAND can be used as a drop-in replacement for the conventional overlapper in
the Celera Assembler, or can be used as a standalone tool by advanced users.
<p>SAND is part of the <a
href=http://www.cse.nd.edu/~ccl/software>Cooperating Computing
Tools</a>. You can download the CCTools from <a
href=http://www.cse.nd.edu/~ccl/software/download>this web page</a>,
follow the <a href=install.html>installation instructions</a>, and you
are ready to go.</p>
<h2>Using SAND with the Celera Assembler</h2>
If you are already using the Celera Assembler version 5.4, you can easily switch to using SAND
to accelerate assemblies by using our modified <tt>sand_runCA_5.4</tt> script. If you are using the newer
Celera Assembler version 6.1, you can switch to SAND by using the <tt> sand_runCA_6.1 </tt> script. If you are using
Celera Assembler version 7.0, you can switch to SAND by using the <tt> sand_runCA_7.0 </tt> script. We assume that you have
installed SAND according to the instructions above and addeed it to your <tt>PATH</tt>.
<ol>
<li> Copy the script <tt>sand_runCA_X.X</tt> into the same directory where you have the normal <tt>runCA</tt> installed.
<li> Set <tt>ovlOverlapper=sand</tt> to your <tt>spec</tt> file.
<li> For Celera 7.0, add <tt>sandAlignFlags=-e "-o ovl_new"</tt> to your <tt>spec</tt> file. This tells the alignment program
to output new Celera 7.0 style overlaps.
<li> Run <tt>sand_runCA_X.X</tt> just like you normally use <tt>runCA</tt>
</ol>
You will see the assembly start as normal. When the overlapping stage begins, you will see output like this:
<pre>
Total | Workers | Tasks Avg | Candidates
Time | Idle Busy | Submit Idle Run Done Time | Found
0 | 0 0 | 0 0 0 0 nan | 0
5 | 0 0 | 0 0 0 0 nan | 0
10 | 0 0 | 0 0 0 0 nan | 0
15 | 0 0 | 0 0 0 0 nan | 0
...
</pre>
SAND is now waiting for you to start <tt>work_queue_worker</tt> processes.
Each worker that you start will connect back to the master process
nd perform small pieces of the work at a time.
In general, the more machines that you can harness, the faster the work will go.
<p>
For testing SAND, just open a new console window and start a single worker,
specifying the hostname where the master runs and the port number it is listening on.
(This can be changed with the sandPort option in the CA spec file.)
For example:
<pre>
% work_queue_worker master.somewhere.edu 9123
</pre>
With one worker, you will see some change in the output, like this:
<pre>
Total | Workers | Tasks Avg | Candidates
Time | Idle Busy | Submit Idle Run Done Time | Found
0 | 0 0 | 0 0 0 0 nan | 0
5 | 0 1 | 100 83 1 16 0.32 | 1858
10 | 0 1 | 100 69 1 30 0.33 | 3649
15 | 0 1 | 100 55 1 44 0.34 | 5464
</pre>
For a very small assembly, a single worker might be sufficient.
However, to run a really large assembly at scale, you will need to
run as many workers as possible. A simple (but tiresome) way of doing
so is to <tt>ssh</tt> into lots of machines and manually run <tt>work_queue_worker</tt>
as above. But, if you have access to a batch system like
<a href=http://www.cs.wisc.edu/condor>Condor</a> or
<a href=http://www.sun.com/software/sge>SGE</a>, you can use them
to start many workers with a single submit command.
<p>
We have provided some simple scripts to make this easy.
For example, to submit 10 workers to your local Condor pool:
<pre>
% condor_submit_workers master.somewhere.edu 9123 10
Submitting job(s)..........
Logging submit event(s)..........
10 job(s) submitted to cluster 298.
</pre>
Or, to submit 10 worker processes to your SGE cluster:
<pre>
% sge_submit_workers master.somewhere.edu 9123 10
Your job 1054781 ("worker.sh") has been submitted
Your job 1054782 ("worker.sh") has been submitted
Your job 1054783 ("worker.sh") has been submitted
...
</pre>
Note that <tt>condor_submit_workers</tt> and <tt>sge_submit_workers</tt>
are simple shell scripts, so you can edit them directly if you would
like to change batch options or other details.
<p>
Once the workers begin running, the SAND modules can dispatch tasks to
each one very quickly. It's ok if a machine running a worker crashes
or is turned off; the work will be silently sent elsewhere to run.
<p>
When the SAND module's master process completes, your workers will
still be available, so you can either run another master with the
same workers, remove them from the batch system, or wait for them to
expire. If you do nothing for 15 minutes, they will automatically exit.
<h2>SAND in More Detail</h2>
This section explains the two SAND modules in more detail,
if you would like to tune the performance or use them independently
of Celera. We assume that you begin with a file of reads in FASTA format.
To use the SAND modules, you must first generate repeats and compress the data. (If you don't have data handy, download <tt>small.cfa</tt> and <tt>small.repeats</tt> data from the <a href=http://www.nd.edu/~ccl/software/sand>SAND Webpage</a>.)
To generate repeats from a FASTA file, use the <tt>meryl</tt> tool from the Celera Assembler:
<pre>
% meryl -B -m 24 -C -L 100 -v -o small.meryl -s small.fa
% meryl -Dt -s small.meryl -n 100 > small.repeats
</pre>
Then use <tt>sand_compress_reads</tt> to compress the sequence data into a compressed FASTA (.cfa) file:
<pre>
% sand_compress_reads small.fa small.cfa
</pre>
The filtering step will read in the compressed sequence data (<tt>small.cfa</tt>) and quickly produce a list of candidate sequences (<tt>small.cand</tt>) for the following step to consider in detail. Start the filtering step as follows:
<pre>
% sand_filter_master -r small.repeats small.cfa small.cand
</pre>
While the filtering step runs, it will print some statistics to the
console, showing the number of workers available, tasks running, and so forth:
<pre>
Total | Workers | Tasks Avg | Candidates
Time | Idle Busy | Submit Idle Run Done Time | Found
0 | 0 0 | 0 0 0 0 nan | 0
5 | 0 14 | 158 14 14 130 0.39 | 16452
10 | 0 15 | 356 15 15 326 0.38 | 42382
15 | 0 15 | 549 15 15 519 0.38 | 69055
20 | 0 15 | 744 15 15 714 0.38 | 96298
25 | 0 15 | 942 15 15 912 0.38 | 124284
</pre>
The alignment step will take the list of candidates generated in the previous step (<tt>small.cand</tt>),
the compressed sequences (<tt>small.cfa</tt>) and produce a listing of how and where the sequences
overlap (<tt>small.ovl</tt>). For example:
<pre>
% sand_align_master sand_align_kernel -e "-q 0.04 -m 40" small.cand small.cfa small.ovl
</pre>
The options <tt>-q 0.04 -m 40</tt> passed to <tt>sand_align_kernel</tt> indicate a minimum alignment quality of 0.04 and a minimum alignment length of 40 bases.
Again, a progress table will be printed to standard out:
<pre>
Total | Workers | Tasks Avg | K-Cand K-Seqs | Total
Time | Idle Busy | Submit Idle Run Done Time | Loaded Loaded | Speedup
0 | 0 0 | 0 0 0 0 0.00 | 0 0 | 0.00
8 | 0 48 | 100 52 48 0 0.00 | 1000 284 | 0.00
10 | 0 86 | 100 13 86 1 7.07 | 1000 284 | 0.71
36 | 1 83 | 181 14 83 2 19.47 | 1810 413 | 1.08
179 | 1 83 | 259 92 83 3 22.51 | 2590 1499 | 0.38
186 | 2 80 | 259 15 80 85 28.54 | 2590 1499 | 13.04
199 | 2 80 | 334 90 80 86 29.96 | 3340 1499 | 12.95
200 | 2 80 | 334 90 80 114 59.43 | 3340 1499 | 33.88
202 | 2 81 | 334 9 81 165 86.08 | 3340 1499 | 70.32
</pre>
After the sequence alignment step completes, you will have an overlap (<tt>.ovl</tt>)
file that can be fed into the final stages of your assembler to complete the consensus step.
<h2>Tuning Suggestions</h2>
<dir>
<li> As a rule of thumb, a single task should take a minute or two. If tasks are much longer than that, it becomes more difficult to measure progress and recover from failures. If tasks are much shorter than that, the overhead of managing the tasks becomes excessive. Use the <tt>-n</tt> parameter to increase or decrease the size of tasks.
<li> When using banded alignment (the default), the <tt>-q</tt> match quality parameter has a significant effect on speed. A higher quality threshhold will consider more alignments, but take longer and produce more output.
<li>
The columns of the output are as follows:
<dir>
<li> Total Time is the elapsed time the master has been running.
<li> Workers Idle is the number of workers that are connected, but do not have a task to run.
<li> Workers Busy is the number of workers that are currently running a task.
<li> Tasks Submitted is the cumulative number of tasks created by the master.
<li> Tasks Idle is the number of tasks waiting for a worker.
<li> Tasks Running is the number of tasks currently running on a worker.
<li> Tasks Done is the cumulative number of tasks completed.
<li> Avg Time is the average time a task takes to run. An average time of 60 seconds is a good goal.
<li> K-Cand Loaded indicates the number of candidates loaded into memory (in thousands).
<li> K-Seqs Loaded indicates the number of sequences loaded into memory (in thousands).
<li> Speedup is the approximate speed of the distributed framework, relative to one processor.
</dir>
</dir>
<h2>For More Information</h2>
For the latest information about SAND, please visit our <a href=http://www.cse.nd.edu/~ccl/software/sand>web site</a> and subscribe to our <a href=http://www.cse.nd.edu/~ccl/software>mailing list</a>.
</body>
</html>
|