These three files, GramNeg, TypGramPos, and AtypGramPos are the same
results as taken from the original file proteomes_full_1.hits
(explained in README.old) and modified as follows:

1. 10 species were erroneously marked of 'Unknown'
gram-classification.  This has been corrected in all 7 files.

2.  The original 'whole protein score' used to find and sort hits was
replaced by a new z-value, and filtered at the more stringent cutoff
of 4.0, which corresponds to a whole protein accuracy of 25%.  That
is, a protein scoring 4.0 z-value has a 25% chance of being a TMB
protein.

In short, z-value is in units of standard deviations.  That is, a
z-value of 4.0 means 4 standard deviations from the average bits score
received from PROFtmb on a set of similar-length non-TMB proteins.
Z-value corrects PROFtmb bits score dependence on protein length.
This dependence is approximately linear, which is why the
linear-adjustment of the whole protein score, (bits-(Length*0.066-6)
was servicable, but Z-value is more accurate, especially for short
proteins.  For more details, please see 'Methods' tab.


The three files GramNeg, TypGramPos, and AtypGramPos have identical
fields except for Field 2, which is changed from Whole protein score
to Z-value


Field 1:	GenBank ID
Field 2:	Z-value
Field 3:	Protein Length
Field 4:	Number of Predicted transmembrane strands
Field 5:	Classification (Gram-, Gram+, or Atypical Gram+)
Field 6:	Organism
Field 7:	Annotation or Homology to Annotation ('---' if none)
Field 8:	4-state per-residue Prediction
Field 9:	Amino Acid Sequence