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
|
RTAX: Rapid and accurate taxonomic classification of short paired-end
sequence reads from the 16S ribosomal RNA gene.
David A. W. Soergel (1), Rob Knight (2), and Steven E. Brenner (1)
1 Department of Plant and Microbial Biology,
University of California, Berkeley
2 Howard Hughes Medical Institute and Department of Chemistry
and Biochemistry, University of Colorado at Boulder
* Corresponding author (current address): soergel@cs.umass.edu
Version 0.984 (August 7, 2013)
http://dev.davidsoergel.com/rtax
Requirements
------------
* Hardware: memory requirements vary with the size of the reference database;
for the version of GreenGenes we provide, ~4GB is needed.
* USEARCH (http://www.drive5.com/usearch/). RTAX has been tested with
v4.1.93 and v5.0.144. Make sure that the program is in your search
path as "usearch", or provide its path explicitly in the "rtax" script.
Note that v5.x uses quite a bit more memory, so 4.x may be preferred.
* A reference database consisting of sequences for which taxonomy assignments
are known. We provide prepared versions of the GreenGenes database on the
RTAX web site; see the GreenGenes section below for details.
Two files are needed:
1. a FASTA file containing the sequences, and
2. a file listing taxonomy strings for each sequence ID found in the
FASTA file. The format of this file is
sequenceid pcidwidth kingdom; phylum; class; order; ...
one entry per line, where sequenceid, pcidwidth, and the taxonomy
string are separated by a single tab, and the taxonomy string itself
is delimited either by semicolons or by tabs. The pcidwidth column
is optional, and (if present) is ignored in the current version of
RTAX anyway. (In our usage, we cluster the reference database into
99% id OTUs, representing each by a single "seed" sequence. This
column then lists the largest pairwise %id difference between the
cluster seed and any cluster member.)
Installation
------------
RTAX consists of a set of perl scripts and a shell script ("rtax") to wire
them together. No compilation is needed; just extract the package to a
convenient location.
The perl scripts must remain in the "scripts" directory below the "rtax" shell
script, but the latter can be symlinked anywhere for convenience. A common
pattern might be to place the whole package at /usr/local/rtax, and then
symlink the script into your path (e.g., "ln -s /usr/local/rtax/rtax
/usr/local/bin").
Running RTAX
------------
Sequence classification is done as follows:
rtax -r gg.nr.fasta -t gg.nr.taxonomy -a queryA.fasta [-b queryB.fasta] -o result.out
Substitute a different reference database for the GreenGenes files if desired,
of course.
If two query files are provided, then they must contain mate-paired reads
with exactly matching identifiers (though they need not be in the same order).
Any ids present in one file but not the other are silently dropped. Alternate
naming schemes for the paired reads (e.g., "SOMEID.a" paired with "SOMEID.b",
and so forth) are handled via the -i option.
RTAX considers query sequences in the forward sense with respect to the
reference database. Thus, in a paired-end experiment, the reverse reads will
likely need be reverse-complemented with the -y option.
RTAX may be run for single-ended reads as well; simply exclude the queryB
parameter in this case.
Run "rtax" with no arguments for a help message describing additional options.
Various progress messages are provided on stderr, and the predicted
classification for each query is provided in the file given by the -o option.
The output format is essentially the same as the taxonomy file input format
above. The second column indicates the best %id observed between the query
sequence and any reference sequence, and the taxonomy string is tab-delimited.
Temporary files are created in /tmp/rtax.SOMENUMBER (or under the temp
directory given by -m). These can be quite large; please use -m to select a
location with sufficient free space, as needed. The temporary directory is
deleted upon successful termination, but may need to be manually cleaned up in
the event of an error of some sort.
Using the GreenGenes reference database
---------------------------------------
An RTAX-formatted reference database based on GreenGenes (version of March 11,
2011) is available at
http://dev.davidsoergel.com/rtax/greengenes.reference.20110311.tgz. That
contains gg.nr.fasta and gg.nr.taxonomy, which are the result of clustering
the GreenGenes input file at 99% identity, finding consensus taxonomy strings
for each cluster, and formatting the result as needed for use with RTAX.
Please see the "greengenes" subdirectory in the RTAX distribution for
information on using newer versions of GreenGenes.
Preparing a non-GreenGenes reference database
---------------------------------------------
You can certainly use different databases, or GreenGenes clustered by
different means or with different parameters (or not at all, though this
approach will have poor performance). If you have any trouble producing the
reference fasta and taxonomy files in the required format, examine the
prepare-greengenes script for hints, and feel free to contact us for help.
Plumbing
--------
Taxonomy assignment proceeds in two phases, which we implement separately:
1. For each query sequence (or mate pair), find an appropriate set of
matching hits in a reference database. This is implemented as
rtaxSearchSingle and rtaxSearchPair for single and paired-end reads,
respectively.
2. Find consensus taxonomy among each set of reference hits. This is
implemented as rtaxVote.
3. Finally the detailed rtaxVote results are filtered and cleaned up
for output.
|
