Contents of directories
-----------------------

hmm - SMART HMMER2 Hidden Markov Models

aln - SMART alignments

consensus - SMART alignment consensus sequences


Linking to SMART annotation
---------------------------

Links of the form:

http://smart.embl.de/smart/do_annotation.pl?ACC=512&BLAST=DUMMY

should be stable. You can also specify accessions as (for instance):

SM00530 - A leading 'sm' or 'SM' is ignored, and the accession treated as 
a number.

Notes on table of thresholds
----------------------------

a) Domain

This is the SMART domain name. This will correspond directly to an entry
on http://smart.embl.de/

The commercial distribution of SMART may also contain pre-release domains that 
are not currently available on the web. The SMART team should be contacted
if further information is required in these cases.

b) Family

Some domains belong to larger families. For example, RAS, RHO, RAB, RAN, ARF
and SAR are all sub families of the larger small_GTPase family. The family
field contains the name of the larger family to which a domain belongs, if
such a family exists. However, for the majority of domains, there are no
sub-families in which case this field will be the same as a).

c) DB size

This is the database size for which significance values (i.e. the thresholds)
were calculated. This is the number that should be used as the argument for
the -Z parameter in hmmsearch. Alternatively, together with mu, lamda (see 
later) and the bit score for a domain match, this value can be used to 
calculate an E-value directly comparable to the cutoffs. (This may be useful
when examining hmmpfam output).

d) mu

HMM mu parameter

e) lambda

HMM lamda parameter.

f) cutoff

This is the E-value of the lowest scoring true positive.

g) cut_low

This is the E-value of the highest scoring false positive.

h) family

Matches more significant than cut_family, but less significant than cutoff 
can be regarded as members of the family classification (for instance 
small_GTPase), but not as being members of an individual sub-family.
Only domains that are members of larger groups have an associated cut_family 
value.

i) repeats

The minimum number of repeats that must be present for this domain to be
regarded as a true hit.

j) repeat_cut

If the minimum number of repeats is present, this value is used as the E-value
threshold for that particular domain

The SMART scoring scheme is complicated, and thoroughly explained in:

Schultz, J., Milpetz, F., Bork, P., and Ponting, C.P. (1998)
"SMART, a simple modular architecture research tool:
 Identification of signalling domains" Proc. Natl. Acad. Sci. USA 95, 5857-5864."

--------------------------------
SMART 01/06/2016 email: smart@embl-heidelberg.de
