Given a FASTA/FASTQ file and a list of barcodes (name and seq),
the FASTA/FASTQ file is split into N number of new files representing
the reads for each barcode. The format of the input file is determined
automatically. The barcode will be removed from read sequence (and quality
score if FASTQ).

barcode input format:
tagname	sequence	orientation (5 or 3)	strip tag? (y/n, optional - default y)

The output files will be named: out_template_tagname.fast[qa]

Reads missing a barcode (or if the tag is too degenerate) will be written to:
out_template_missing.fast[qa]

Note: This is a slow method using a Smith-Waterman alignment algorithm. It is
      primarily useful with data that is inherently noisy, such as PacBio
      sequencing reads. For less-noisy sequences, a faster, non-alignment
      approach will be better.

Note 2: This isn't appropriate for color-space FASTQ files with a prefix base
        included in the read sequence, since it trims an equal number of bases
        from the sequence and quality FASTQ lines.


Usage: barcode_split.py {options} barcodes.txt input.fastx{.gz} output_template

Options:
  -edit num         Number of mismatches/indels to allow in barcode discovery.
                    (default: 0)

  -pos num          Allow barcode to be within [num] bases from the start.
                    (or end for 3' barcodes) (default: 0)

  -allow-revcomp    Allow barcodes to match the reverse-compliment of a read.
                    (non-strand specific sequencing) The read's orientation
                    will *not* be changed in the output file.

  -gz               GZip compress the output files

  -stats            Output stats file (output_template.stats.txt)