RFAM version 7.0 and EMBL version r88 was used to construct the benchmark. All sequence datasets 
are in FASTA format.

Each file in the benchmark begins with an RNA family name, a unique number of 
the alignment, and a file suffix indicating the contents. 

------------------

Benchmark without query RNA flanks:

The general format of files in this distribution is

Filename format
<RNA>.<n>.flankunaligned	The query RNA and the genomic sequence containing the target, both 
				unaligned without gaps
<RNA>.<n>.flankaligned		The query RNA aligned to the target RNA in the genomic sequence 
				surrounded by flanking endgaps.
<RNA>.<n>.seq			The genomic sequence of	the pair above, without gaps
<RNA>.<n>.startend		The start and end indices of the target RNA embedded in the 
				genomic sequence. Indices start at 0 and end at length-1.
-------------------

Benchmark with query RNA flanks:

The general format of files in this distribution is

Filename format
<RNA>.<n>.gengen5		The query RNA, with 5' and 3' flanks of 50 nucleotides, 
				and the genomic sequence containing the target, both 
				unaligned without gaps
<RNA>.<n>.query_startend5	The start and end indices of the query RNA embedded in the 
				flanks. Indices start at 0 and end at length-1.

<RNA>.<n>.gengen4		The query RNA, with 5' and 3' flanks of 100 nucleotides, 
				and the genomic sequence containing the target, both 
				unaligned without gaps
<RNA>.<n>.query_startend4	The start and end indices of the query RNA embedded in the 
				flanks. Indices start at 0 and end at length-1.

<RNA>.<n>.gengen6		The query RNA, with 5' and 3' flanks of 150 nucleotides, 
				and the genomic sequence containing the target, both 
				unaligned without gaps
<RNA>.<n>.query_startend6	The start and end indices of the query RNA embedded in the 
				flanks. Indices start at 0 and end at length-1.
---------------------

False positive set:

The general format of files in this distribution is

Filename format
<RNA>.<n>.flankunaligned	The query RNA and a genomic sequence containing a target RNA from a 
				different, both	unaligned without gaps
<RNA>.<n>.gengen5		The query RNA, with 5' and 3' genomic flanks of 50 nucleotides,
				and a genomic sequence containing a target RNA from a 
				different family, both unaligned without gaps
<RNA>.<n>.gengen4		As above but query flanks of 100 nucleotides
<RNA>.<n>.gengen6		As above but query flanks of 150 nucleotides
-------------------

The RFAM_subsets.tgz file contains RFAM family alignments from which the benchmark was
created. Each alignment is at most 50 sequences which were randomly selected from the
original seed alignments which can contain as many as a 1000 sequences. Each file in
RFAM_fixed is of the format

<RNA>.unaligned			The RNA sequences in unaligned format.
<RNA>.aligned			The RNA sequences in aligned format.