<-- README -->

README for NCBI Microbial Genome Submission Check Tool

This is a DEVELOPMENT/BETA project and will undergo numerous revisions.
_________________________________________________________________________

       
       National Center for Biotechnology Information (NCBI)
             National Library of Medicine
             National Institutes of Health
             8600 Rockville Pike
             Bethesda, MD 20894, USA
             tel: (301) 496-2475
             fax: (301) 480-9241
             email: info@ncbi.nlm.nih.gov (for general questions)
             email: genomes@ncbi.nlm.nih.gov (for specific questions)
             
_________________________________________________________________________

=========================================================================


Microbial Genome Submission Check Tool

Command line version

Microbial Genome Submission Check Tool (subcheck) is for the validation of
genome records prior to submission to GenBank. It utilizes a series of
self-consistency checks as well as comparison of submitted annotations to
computed annotations. Some of specified computed annotations could be
pre-computed using BLAST and its modifications and tRNAscanSE. Currently
there is no specific tool for predicting rRNA annotations. Please use the
format specified in documentation

DOWNLOAD

Download the tarballs from FTP location ftp:/ftp.ncbi.nlm.nih.gov/...

INSTALL

Copy the downloaded tarballs to the directory you wish to install the tool.
Change directory to the installation directory and install the tool by typing
the command:

tar xvzf <tarball_name>

for the chosen tarballs.

Format of command line:

<full_path>/subcheck [-h] [-help] -in input_asn [-out output_asn]
    [-inblast blast_res_proteins] [-inblastcdd blast_res_cdd]
    [-intrna input_trna] [-inrrna input_rrna]
    [-parentacc parent_genome_accession] [-inparents InputParentsFile]

REQUIRED ARGUMENTS
 -in <File_In>
   input file in the ASN.1 format, must be either Seq-entry or Seq-submit

OPTIONAL ARGUMENTS
 -h
   Print USAGE and DESCRIPTION;  ignore other arguments
 -help
   Print USAGE, DESCRIPTION and ARGUMENTS description;  ignore other arguments
 -out <File_Out>
   output file in the ASN.1 format, of the same type (Seq-entry or Seq-submit)
 -inblast <File_In>
   input file which contains the standard BLAST output results (ran with -IT
   option) for all query proteins sequences specified in the input genome
   against a protein database (recommended: bact_prot database of Refseq
   proteins supplied with the distributed standalone version of this tool)
 -inblastcdd <File_In>
   input file which contains the standard BLAST output results for all query
   proteins sequences specified in input_asn against the CDD database
 -intrna <File_In>
   input tRNAscan predictions in default output format, default value is <-in
   parameter>.nfsa.tRNA
 -inrrna <File_In>
   input ribosomal RNA predictions (5S, 16S, 23S), see the manual for format,
   default value is <-in parameter>.nfsa.rRNA
 -parentacc <String>
   Refseq accession of the genome which protein annotations need to be
   excluded from BLAST output results
 -inparents <File_In>
   contains a list of all protein accessions/GIs for each Refseq accession/GI

RESULTS

The script creates several files with the results. Most of the file names are based on 
the input file name (input_asn):

<input_asn>.<problem_type>.problems.log

where "problem_type" is one of the seven problem types:
complete.overlap - Completely overlapping genes (on both strands) are reported. 
  Annotations of this type are suspicious and should be manually inspected.
overlap - Partial overlaps are any type of overlap above the distance of 30 
  bases. Overlaps of this type do occur in genome annotations and are reported 
  here for manual inspection. Many overlaps or overlaps of significant length 
  might indicate annotation problems.
rna.overlap - Overlaps between RNA and other features (RNA and CDS) are 
  reported. Annotations of this type are rare and should be manually inspected.
frameshifts - Adjacent genes on the same strand are analyzed for hits against 
  the same subject (common BLAST hit) by comparing BLAST results. Since gene 
  fusions/splits occur in prokaryotic genes the BLAST hits are analyzed for any 
  subject (not the common BLAST hit) that covers 90% of the query protein, in 
  which case the frameshift is not reported under the assumption that this gene 
  is "real". Any pair of genes failing to meet that criteria are reported as 
  potential frameshifts and should be manually inspected.
partial - The results from the RPS-BLAST against conserved domains are analyzed 
  for situations when the conserved domain is only partially covered by the 
  query. Potential truncations will be reported (domain must cover at least 90% 
  of the protein, whereas protein covers 80% or less of the domain). The results 
  should be manually inspected for incorrectly annotated start sites (N-terminal 
  truncation), frameshifts (C-terminal truncations), or any other type of 
  potential problem.
tRNA.missing - RNA annotations are checked with tRNAscanSE for tRNAs, and an 
  internal ribosomal RNA database for structural RNAs. Completely missing tRNAs 
  above a score of 60.0 and missing high-scoring ribosomal RNAs are reported. 
  These should be examined to see if they can be added to the genome. This 
  extension (and some others) is kept for historical reasons and will be changed 
  to "RNA.missing" in a future revision.
tRNA.bad.strand - Occasionally RNAs are incorrectly annotated on the wrong 
  strand (which is difficult to do with protein coding genes). When overlapping 
  computed and submitted RNA annotations have opposite strands they are reported 
  and should be examined to see if the incorrect strand was used in the genome 
  annotation.

The formats of those files are described in MANUAL.standalone

OPTIONAL REQUIREMENTS

You might want to use following tools to generate input files:

tbl2asn - often used for genome submission.

https://www.ncbi.nlm.nih.gov/Genbank/tbl2asn2.html

asn2all - to extract nucleotide and protein sequences from the input ASN.1 file. 

ftp://ftp.ncbi.nih.gov/asn1-converters/by_program/asn2all/

  Recommended command-line parameters:

  To generate nucleotide sequence:

  asn2all -ff -i <input_sqn> -o <nucleotide_fasta_file>

  To generate protein sequences:

  asn2all -ff -i <input_sqn> -v <proteins_fasta_file>

BLAST - either standalone or web-based tool, both available at NCBI - to run 
  protein sequences against a database of existing proteins. We recommend to 
  use the refseq_protein database on the website or more rigorous database on our
  FTP site (ftp://ftp.ncbi.nlm.nih.gov/genomes/Bacteria/all.faa.tar.gz). Instructions
  for creating BLAST databases can be found in the BLAST documents. BLAST output
  must be formatted with the -IT option in order to have subcheck read the results. 
  For our internal subcheck calculations we use the following parameters:
  -FT -IT -e 1e-06 -z 500000000

RPSBLAST - either standalone (rpsblast) or web-based 
  (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi) tool, both available at 
  NCBI - to run protein sequences against CDD (conserved domain database: 
  ftp://ftp.ncbi.nih.gov/pub/mmdb/cdd). 

tRNAscan-SE - tRNA prediction tool (http://lowelab.ucsc.edu/tRNAscan-SE/) - to
  predict tRNA locations given your nucleotide sequence.

cmsearch - RNA prediction tool (https://www.sanger.ac.uk/Software/Rfam/) - to
  predict 5S rRNA locations given your nucleotide sequence. The INFERNAL software
  package contains cmsearch. (http://infernal.janelia.org/).

You can predict 16S and 23S rRNAs by running BLAST of your nucleotide sequence 
against databases in our package in the lib/ directory (RibosomXS, where X is 
16S or 23S) and format the output in a format described in MANUAL.standalone

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

Microbial Genome Submission Check Tool runs on Linux computers.


Edited. Jan 15, 2008. NCBI


<-- End of File -->