>DJS045A03F template=DJS054A03 dir=F library=DJS045 trim=12-543

The data analysis software looks for the six “name=” strings on the description line, and then takes the text from the “=” to the next whitespace character as the “value” of the annotation string. So, other text besides annotation strings can appear on the description line, and no whitespace may appear in the value of an annotation string.

The first three annotation strings, template, dir and library, are used to specify the Paired End information for a read; only reads with template, dir and library annotations will be treated as Paired End reads. The data analysis software performs exact string matching of the template and library annotations to determine which reads come from the same template or library. Multiple reads may have the same template, dir and library information: they all will be assembled or mapped but only the best aligned sequence (by aligned length or number of differences from the consensus) will be used further. There is no requirement that the accession number for the read encode the Paired End information, and the data analysis software will not try to “parse” the accession number for any information. However, if you use a naming convention for encoding the Paired End information in the accession numbers for reads, the fnafile command can be very useful for translating that encoded information into the annotation strings that the data analysis software can then read. (See Section 3.4 for a description of this use.)

The trim annotation describes the region of the sequence that the data analysis software should use for assembly or mapping (with the assumption that the rest of the sequence is non-sample or low quality). For a read 800 bases in length and a trim annotation string “trim=12-543”, for example, the assembler will ignore bases 1-11 and 544-800, and only use bases 12-543 in the assembly or mapping. If either number in the trim annotation string is 0, the beginning or end of the read will be treated as the trimming point (i.e., for the same example read as above, “trim=12-0” would tell the assembler to use bases 12-800 of the read in the assembly).

This trim annotation should combine the results of all of the sources of trimming that may occur (low quality, vector, primer, adapter, linker, etc.), so that the data analysis software is given just the sequence region that represents the bases to be included in the assembly or mapping. The “‑v” option can be used with the runAssembly, runMapping and runProject commands to specify a FASTA file containing sequences to be trimmed: each read will then be screened against this database, and the ends of reads that match the sequences included will be trimmed off.

The scf and phd annotations provide a link back to the “trace” information for the read, and are only used when generating the full consed directory structure (when the ‑consed option is used). The value string for scf or phd annotations can take one of two forms, either as a simple path to the SCF or PHD file, or as a “command string” to be executed to access the SCF or PHD file contents. If the value string does not begin with “cmd:”, it is treated as a path. If the value string begins with “cmd:”, it should have the format “cmd:commandline”, where any whitespace in the commandline portion of the command should be replaced with colons.