AGP Validation

AGP file structure and content can be validated using the agp_validate program. The on-line version of agp_validate checks that the input AGP file conforms to the AGP format Specification, checks the file for internal consistency, and generates a report of component, gap, scaffold and object statistics. The command-line version of agp_validate performs all the same checks as the on-line version but also has options to perform additional checks by comparing the AGP to sequences in FASTA files (see below).

AGP validation on-line

The on-line AGP Validation form can be used to run agp_validate on an uploaded AGP file or on AGP text pasted into the form.

agp_validate command-line program

agp_validate is available by anonymous FTP. Copy the appropriate version for your platform, then uncompress the file, rename it to "agp_validate", and set the "execute" permission (see platform-specific details).

usage overview: agp_validate [-options] [FASTA files...] [AGP files...]

• Run without any options agp_validate will perform a large number of validations on the input AGP files (see below), and will also generate a report of component, gap, scaffold and object statistics.
• If component FASTA sequence files are provided, agp_validate will also check that component spans do not exceed the sequence length.
• If the component sequences are available in GenBank, then agp_validate can perform additional checks using the sequence lengths, versions, and taxonomy ID retrieved from GenBank (-alt and -species options).
• If FASTA sequences for the assembled objects are provided, agp_validate can also check that the sequences match what can be constructed from the AGP and the component sequences (-comp option).
• Information on all the available options can be obtained by executing agp_validate with the -help option.

Validations performed by agp_validate

Error level violations reported include:

• Incorrect number of columns: there should be 9 tab-separated columns.
• Non-positive integers in the following columns:
  • 2: object_beg
  • 3: object_end
  • 4: part_number
  • 6b: gap_length
  • 7a: component_beg
  • 8a: component_end
• object_end is less than the object_beg.
• component_end is less than the component_beg.
• The length of the span specified for the component (in column 7a and 8a) does not match the length of the span specified for the object (in column 2 and 3).
• The length specified for the gap (in column 6b) does not match the length of the span specified for the object (in column 2 and 3).
• Linkage=yes with a gap_type other than scaffold or repeat.
• Object does not start with an object_beg coordinate of 1.
• Object has ranges that are non-sequential and/or overlapping.
• Object does not start with a part_number of 1.
• Object has non-sequential lines and/or lines mixed with other objects.
• Multiple objects with the same object name (column 1).
• Component orientation of 0 or na used for a non-singleton scaffold.
• Invalid terms or symbols in the following columns:
  • 5: component_type
  • 7b: gap_type
  • 8b: linkage
  • 9a: orientation

Warning level violations reported include:

• Gap at the beginning or the end of an object.
• Consecutive gap lines of the same type.
• Overlapping spans used for a given component_id.
• Non-draft component_id used more than once.
• Non-draft component spans out of order.
• Extra tab character at the end of the line.
• Component type is not consistent with the line format.
• Component type is not consistent with the component_id accession.

Additional errors and warnings reported when optional validations are invoked:

• Invalid component_id. [-alt or -g option]
• Component is not in GenBank. [-alt option]
• component_id is ambiguous without an explicit version. [-alt option]
• component_end is greater than the sequence length. [-alt option, or FASTA files provided]