Global parameters
For more explanations, please look at the examples
Title
CaspR need a short title (5 Char max).
It will be used for your sequence, the models...
This name is more convenient than the job number (14 characters) and will allow the user to differenciate between two jobs
Molecule(s) per asymmetric unit
This is the number of identical molecules in the asymetric unit.
Soon CaspR will be able to handle up to 3 different substructures (for crystals containing hetero-multimers) however the computing time will be tremendously increased.
Reflection file (in MTZ format)
CaspR is expecting a truncated MTZ reflection file (CCP4 format). It will thus automatically extract the cell parameters, space group and flags for F and SIGF("mtzdmp" command).

CaspR parameters
For more explanations, please look at the examples
Input fasta file
This file is one of the most important input for the CaspR process, so be cautious. The first sequence should correspond to your protein sequence in FASTA format. You can also provide some related sequences in order to increase the overall multiple alignment quality. (i.e. 2 to 7 related sequences sampling the evolutionary divergence of the family and linking together the provided target sequence with the provided structure(s))
But please do not enter more than 10 sequences ...
The sequence must respect the standard one letter code (special residues notations B,U,Z,X,*,- are not allowed), each sequence should have a different name

EXAMPLE:
>yoursequence
MKIKAVGPLLCAGITTYSPLRHWQA
ALPISYRTHHHGAFDEP
>RELATED_SEQ_1
MKHHAVGPLLCAGITTYSPLRHWQA
ALPISYRTPLLCAGITTYSHHHGAFDEP
>RELATED_SEQ_2
MLPKAVGPLLCAGITTYSPLRHWQA
ALIKAVGPLLPISYRTHHHGAFDEP
...
Structure(s)
Enter here at least one structure to be used for the alignment.
If there is more than one structure, the program T-COFFEE/SAP will be used to generate a structural alignment.
You can enter the structure(s) as a PDB code or upload your PDB file.
(PDB format: ATOM 1 CB GLN 45 52.568 -3.205 -8.078 1.00 68.55).
PDB code : with the PDB code you should provide the chain name. Leave the chain field empty if the chain has no identifier (blank). If you enter the wrong chain name, the program will crash.
When the provided PDB code is absent from our local database, a WARNING will be displayed, you should then upload the file.
Uploaded file : You can upload up to 3 PDB files. If your file contain several chains, the first chain will be used.
Email adress
Enter here a valid Email adress
A first mail is sent when the job is submitted and a second mail upon job completion. It is always possible to consult your job status by inputing the proper job number in the the "job number" field.

Results
For more explanations, please look at the examples
Status
You can find here general information about your job :
-date of submission
- email address
- name of your molecule
- number of molecules per asymmetric unit
- names of your uploaded files
- structure(s) used by CaspR
Read MTZ File
Data such as cell dimensions, space group, resolution and number of reflections are directly extracted from the MTZ file.
CaspR also read the various space group format recognized by the different programs used in the procedure.
Since they is no manual check on your data you have to verify there is no mistakes.
T-COFFEE
Information about the multiple alignment :
-the alignment is produce in HTML and PDF format and gives the CORE index for each position in the alignment using a color code going from blue (bad) to red (good).
-CaspR returns to the user the percentage of identity between your sequence and each of the structure(s) sequences.
-the average score obtained in the multiple alignment (can vary from 0 to 99, the higher, the better) This correspond to the average CORE index associated with every sequence.
Truncation details :
First, CaspR choose a cutoff threshold between 1 and 8, mostly depending on the average score.
Then the program parse the CORE index obtained from the multiple alignment. In this index, each aligned residue is associated with a score between 0 and 9 (the higher, the better).
A residue in the user target sequence is truncated :
- if the CORE index value at that position is below the threshold
- if it is not aligned with any of the structures
- if segment truncations in an area leaves a segment of less than 3 residues length alone, then the latter segment is removed (leading to a longer truncation)
Conversely, if the fragment that is to be truncated is less than 3 residues length , the truncation is cancelled.
The number of truncated segments as well as their positions are displayed on the figure.
MODELLER
MODELLER uses the provided structures to generate 30 models based on the T-COFFEE multiple alignment. The number of models already generated is indicated on top of the MODELLER frame. When all 30 models are generated, the truncations are applied to all models according the T-COFFEE CORE index.
Model names always start with 'm', followed by a name (the title you provided) and 4 digits.
Truncated model names always start with 'Trm', followed by a name (the title you provided) and 4 digits.
A "molmol" representation of the models variability is presented on the PNG figures.
The models are divided into 2 groups, truncated and not truncated models with for each group 2 view modes :
- the 'sausage' mode
- the 'stick' mode
In the 'sausage' mode, all structures are superimposed and an average structure is computed. The average RMSD between the mean structure and the models at a given position is illustrated by the diameter of the ribbon on the figure. The excised segments on the truncated models are coloured in red.
In the 'stick' mode, all structures are superimposed and presented on the figure using a cylinder representation of the bonds.
AMoRe
When the MTZ file is converted into an ascii file, CaspR returns information about the provided data (this is the sort step in AMoRe):
-number of reflections and completeness in the global resolution range
-number of reflections in the resolution range used by AMoRe
A table shows which model has obtained the best score in each of the search steps (rotation, translation, fitting).
AMoRe uses a rotation function, named 'o1r', followed by n translation steps, where n is equal to the number of molecules per asymmetric unit. ('ot1' to 'otn'). Finally, AMoRe proceeds with a rigid body step, named 'ofn (fitting').
A 'C/R' value is computed by dividing the Correlation by the R-factor (C and R in the table), which is used to sort in decreasing order the various models solutions.
Click the button 'AMoRe' to get a graphic view of all models solution.
The number given on top of the summary table allows the user to know the number of models which have already been through the AMoRe
CNS
The 10 best AMoRe model solutions are pre-refined using CNS as well as the structures provided by the user as templates if they are not part of the 10 first ranking solutions in AMoRe. The table on the CNS summary frame only shows the best scoring model after CNS refinement.
CNS, takes the AMoRe solution (initial model) to run one cycle of pre-refinement involving rigid-body and minimization refinement. The pre-refined PDB file is then activated under 'final model' on the CNS frame. The used resolution in CNS is the same than the AMoRe one ( 15 to 3 A).
To visualize the full CNS output, the user should activate the 'CNS' link. The user can then download both un-refined (left link) and pre-refined (right link) models by activating the link on the models name.