.. _fold_assignment_methods:

Fold assignment methods
***********************

ModPipe includes a number of methods for assigning folds and detecting
possible templates for the target sequence. Most of these are implemented by
Modeller, although they can also be supplemented with PSI-BLAST searches.
Typically during a ModPipe run templates are searched for using multiple
methods, as no one method can be guaranteed to find every hit.

Each fold assignment method relies on one or more
:ref:`databases <databases>`.

.. _seq_seq_search:

Sequence-Sequence search
------------------------

Modeller's :meth:`Profile.build` method is used to search for the target
sequence against all sequences in the template structure database (typically
this is PDB95).
        
.. _prf_seq_search:

Profile-Sequence search
-----------------------

The target sequence is scanned against a non-redundant sequence database
(typically UniProt90) to create a target profile. This can be done with
Modeller's :meth:`Profile.build`, PSI-BLAST, or both. This profile is then
scanned against all sequences in the template structure database (usually PDB95)
to detect hits.

.. _prf_prf_search:

Profile-Profile search
----------------------

A target profile is derived from the target sequence as for
:ref:`profile-sequence search <prf_seq_search>`. It is then scanned against
all profiles in the template profile database to find hits.

.. _seq_prf_search:

Sequence-Profile search
-----------------------

The target sequence is scanned against every profile in the template profile
database. Aside from the regular scoring, two additional scoring mechanisms
are available. In Max-PSSM Consensus-Profile scoring, the substitution score
for each target position is derived by finding the residue with the 
highest Position-Specific-Scoring-Matrix value (of the target profile) at that
position, then the substitution score for that residue is read from the
template profile. In Max-frequency Consensus-Profile, the substitution score
for each target position is derived by finding the residue with the
highest weighted frequency (of the target profile) at that position,
then the substitution score for that residue is read from the template profile.