You are reading the documentation for version 2.2 of OpenStructure. You may also want to read the documentation for:
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.7.1
1.8
1.9
1.10
1.11
2.0
2.1
devel

Parameters: 


Returns:  The merged alignment 
Return type: 
Merge a list of pairwise alignments into a multiple sequence alignments. This function uses the reference sequence as the anchor and inserts gaps where needed. This is also known as the star method.
The resulting multiple sequence alignment provides a simple way to map between residues of pairwise alignments, e.g. to compare distances in two structural templates.
There are a few things to keep in mind when using this function:
The reference sequence mustn’t contain any gaps
The first sequence of each pairwise alignments corresponds to the reference sequence. Apart from the presence of gaps, these two sequences must be completely identical.
If the reference sequence has an offset, the first sequence of each pairwise alignment must have the same offset. This offset is inherited by the first sequence of the final output alignment.
The resulting multiple sequence alignment is by no means optimal. For better results, consider using a multiplesequence alignment program such as MUSCLE or ClustalW.
Residues in columns where the reference sequence has gaps should not be considered as aligned. There is no information in the pairwise alignment to guide the merging, the result is undefined.
Example:
ref_seq = ost.seq.CreateSequence('ref', 'acdefghiklmn') seq_a1 = seq.CreateSequence('A1', 'acdefghiklmn') seq_a2 = seq.CreateSequence('A2', 'atdfghikllmn') seq_b1 = seq.CreateSequence('B1', 'acdefghiklmn') seq_b2 = seq.CreateSequence('B2', 'acdqhirlmn') aln_a = seq.CreateAlignment() aln_a.AddSequence(seq_a1) aln_a.AddSequence(seq_a2) print(aln_a) # >>> A1 acdefghiklmn # >>> A2 atdfghikllmn aln_b = seq.CreateAlignment() aln_b.AddSequence(seq_b1) aln_b.AddSequence(seq_b2) print(aln_b) # >>> B1 acdefghiklmn # >>> B2 acdqhirlmn aln_list = ost.seq.AlignmentList() aln_list.append(aln_a) aln_list.append(aln_b) merged_aln = ost.seq.alg.MergePairwiseAlignments(aln_list, ref_seq) print(merged_aln) # >>> ref acdefghiklmn # >>> A2 atdfghikllmn # >>> B2 acdqhirlmn
ValidateSEQRESAlignment
(aln, chain=None)¶Checks if sequence in alignment has same connectivity as residues in chain. This looks for connected stretches in both the sequence and the chain and returns False if they don’t match. This uses the connectivity of the protein backbone.
Parameters: 


Returns:  True if all residues (beside gapped ones) are connected, False otherwise. 
AlignToSEQRES
(chain, seqres, try_resnum_first=False, validate=True)¶Aligns the residues of chain to the SEQRES sequence, inserting gaps where needed. The function uses the connectivity of the protein backbone to find consecutive peptide fragments. These fragments are then aligned to the SEQRES sequence.
All the nonligand, peptidelinking residues of the chain must be listed in SEQRES. If there are any additional residues in the chain, the function raises a ValueError.
Parameters: 


Returns:  The alignment of the residues in the chain and the SEQRES entries. 
Return type: 
AlignmentFromChainView
(chain, handle_seq_name='handle', view_seq_name='view')¶Creates and returns the sequence alignment of the given chain view to the chain handle. The alignment contains two sequences, the first containing all nonligand peptidelinking residues, the second containing all nonligand peptidelinking residues that are part of the view.
Parameters: 


Returns:  The alignment 
Return type: 
Conservation
(aln, assign=true, prop_name="cons", ignore_gap=false)¶Calculates conservation scores for each column in the alignment, according to the ConSurf method (Armon et al., J. Mol. Biol. (2001) 307, 447463).
The conservation score is a value between 0 and 1. The bigger the number the more conserved the aligned residues are.
Parameters: 


LocalAlign
(seq1, seq2, subst_weight, gap_open=5, gap_ext=2)¶Performs a Smith/Waterman local alignment of seq1 and seq2 and returns the bestscoring alignments as a list of pairwise alignments.
Example:
seq_a = seq.CreateSequence('A', 'acdefghiklmn')
seq_b = seq.CreateSequence('B', 'acdhiklmn')
alns = seq.alg.LocalAlign(seq_a, seq_b, seq.alg.BLOSUM62)
print(alns[0].ToString(80))
# >>> A acdefghiklmn
# >>> B acdhiklmn
Parameters: 


Returns:  A list of bestscoring, nonoverlapping alignments of seq1 and seq2. Since alignments always start with a replacement, the start is stored in the sequence offset of the two sequences. 
GlobalAlign
(seq1, seq2, subst_weight, gap_open=5, gap_ext=2)¶Performs a Needleman/Wunsch global alignment of seq1 and seq2 and returns the bestscoring alignment.
Example:
seq_a = seq.CreateSequence('A', 'acdefghiklmn')
seq_b = seq.CreateSequence('B', 'acdhiklmn')
alns = seq.alg.GlobalAlign(seq_a, seq_b, seq.alg.BLOSUM62)
print(alns[0].ToString(80))
# >>> A acdefghiklmn
# >>> B acdhiklmn
Parameters: 


Returns:  Bestscoring alignment of seq1 and seq2. 
ShannonEntropy
(aln, ignore_gaps=True)¶Returns the percolumn Shannon entropies of the alignment. The entropy describes how conserved a certain column in the alignment is. The higher the entropy is, the less conserved the column. For a column with no amino aids, the entropy value is set to NAN.
Parameters: 


Returns:  List of column entropies 
SemiGlobalAlign
(seq1, seq2, subst_weight, gap_open=5, gap_ext=2)¶Performs a semiglobal alignment of seq1 and seq2 and returns the best scoring alignment. The algorithm is Needleman/Wunsch same as GlobalAlign, but without any gap penalty for starting or ending gaps. This is prefereble whenever one of the sequences is significantly shorted than the other. This make it also suitable for fragment assembly.
Example:
seq_a = seq.CreateSequence('A', 'abcdefghijklmnok')
seq_b = seq.CreateSequence('B', 'cdehijk')
alns = seq.alg.GlobalAlign(seq_a, seq_b, seq.alg.BLOSUM62)
print(alns[0].ToString(80))
# >>> A abcdefghijklmnok
# >>> B cdehijk
alns = seq.alg.SemiGlobalAlign(seq_a, seq_b, seq.alg.BLOSUM62)
print(alns[0].ToString(80))
# >>> A abcdefghijklmnok
# >>> B cdehijk
Parameters: 


Returns:  bestscoring alignment of seq1 and seq2. 
Renumber
(seq_handle, sequence_number_with_attached_view=1, old_number_label=None)¶Function to renumber an entity according to an alignment between the model sequence and the fulllength target sequence. The aligned model sequence or the alignment itself with an attached view needs to be provided. Upon succcess, the renumbered entity is returned. If an alignment is given, the first sequence of the alignment is considered the fulllength sequence and it must match the model sequence wherever it is aligned (i.e. excluding gaps).
from ost.seq.alg import renumber
from ost.bindings.clustalw import *
ent = io.LoadPDB("path_to_model")
s = io.LoadSequence("path_to_full_length_fasta_seqeunce")
pdb_seq = seq.SequenceFromChain("model", ent.chains[0])
aln = ClustalW(s, pdb_seq)
aln.AttachView(1, ent.chains[0].Select(""))
e = Renumber(aln.sequences[1])
io.SavePDB(e, "renum.pdb")
Parameters: 


Raises: 

This is a set of functions for predicting pairwise contacts from a multiple sequence alignment (MSA). The core method here is mutual information which uses coevolution to predict contacts. Mutual information is complemented by two other methods which score pairs of columns of a MSA from the likelyhood of certain amino acid pairs to form contacts (statistical potential) and the likelyhood of finding certain substitutions of aminioacid pairs in columns of the MSA corresponding to interacting residues.
ContactPredictionScoreResult
¶Object containing the results form a contact prediction.
matrix
¶An NxN FloatMatrix
where N is the length of the alignment.
The element i,j corresponds to the score of the corresponding
columns of the MSA. High scores correspond to high likelyhood of
a contact.
sorted_indices
¶List of all indices pairs i,j, containing (N*N1)/2 elements, as the matrix is symmetrical and elements in the diagonal are ignored. The indices are sorted from the pair most likely to form a contact to the least likely one.
GetScore
(i, j)¶returns matrix(i,j)
Parameters: 


SetScore
(i, j, score)¶Sets matrix(i,j) to score
Parameters: 


PredictContacts
(ali)¶Predicts contacts from a multiple sequence alignment using a combination of Mutual Information (MI) and the Contact Substitution Score (CoEvoSc). MI is calculated with the APC and small number corrections as well as with a transformation into Zscores. The CoEvoSc is calculated using the default PairSubstWeightMatrix (see seq.alg.LoadDefaultPairSubstWeightMatrix). The final score for a pair of columns (i,j) of ali is obtained from:
Sc(i,j)=MI(i,j)exp(CoEvoSc(i,j)) if (i,j) >=0
Sc(i,j)=MI(i,j)exp(1CoEvoSc(i,j)) if (i,j) <0
Parameters:  ali (AlignmentHandle ) – The multiple sequence alignment 

CalculateMutualInformation
(aln, weights=LoadConstantContactWeightMatrix(), apc_correction=true, zpx_transformation=true, small_number_correction=0.05)¶Calculates the mutual information (MI) from a multiple sequence alignemnt. Contributions of each pair of aminoacids are weighted using the matrix weights (weighted mutual information). The average product correction (apc_correction) correction and transformation into Zscores (zpx_transofrmation) increase prediciton accuracy by reducing the effect of phylogeny and other noise sources. The small number correction reduces noise for alignments with small number of sequences of low diversity.
Parameters: 


CalculateContactScore
(aln, weights=LoadDefaultContactWeightMatrix())¶Calculates the Contact Score (CoSc) from a multiple sequence alignment. For each pair of residues (i,j) (pair of columns in the MSA), CoSc(i,j) is the average over the values of the weights corresponding to the amino acid pairs in the columns.
Parameters: 


CalculateContactSubstitutionScore
(aln, ref_seq_index=0, weights=LoadDefaultPairSubstWeightMatrix())¶Calculates the Contact Substitution Score (CoEvoSc) from a multiple sequence alignment. For each pair of residues (i,j) (pair of columns in the MSA), CoEvoSc(i,j) is the average over the values of the weights corresponding to substituting the amino acid pair in the reference sequence (given by ref_seq_index) with all other pairs in columns (i,j) of the aln.
Parameters: 


LoadDefaultContactWeightMatrix
()¶Returns:  CPE, a ContactWeightMatrix that was calculated from a large (>15000) set of
high quality crystal structures as CPE=log(CF(a,b)/NCF(a,b)) and then normalised so that all its elements are comprised between 0 and 1. CF(a,b) is the frequency of amino acids a and b for pairs of contacting residues and NCF(a,b) is the frequency of amino acids a and b for pairs of noncontacting residues. Apart from weights for the standard amino acids, this matrix gives a weight of 0 to all pairs for which at least one aminoacid is a gap. 

LoadConstantContactWeightMatrix
()¶Returns:  A ContactWeightMatrix . This matrix gives a weight of one to all pairs of
standard aminoacids and a weight of 0 to pairs for which at least one aminoacid is a gap. 

LoadDefaultPairSubstWeightMatrix
()¶Returns:  CRPE, a PairSubstWeightMatrix that was calculated from a large (>15000) set of
high quality crystal structures as CRPE=log(CRF(ab>cd)/NCRF(ab>cd)) and then normalised so that all its elements are comprised between 0 and 1. CRF(ab>cd) is the frequency of replacement of a pair of amino acids a and b by a pair c and d in columns of the MSA corresponding to contacting residues and NCRF(ab>cd) is the frequency of replacement of a pair of amino acids a and b by a pair c and d in columns of the MSA corresponding to noncontacting residues. Apart from weights for the standard amino acids, this matrix gives a weight of 0 to all pair substitutions for which at least one aminoacid is a gap. 

PairSubstWeightMatrix
(weights, aa_list)¶This class is used to associate a weight to any substitution from one aminoacid pair (a,b) to any other pair (c,d).
weights
¶A FloatMatrix4
of size NxNxNxN, where N=len(aa_list)
aa_list
¶A CharList
of one letter codes of the amino acids for which weights are found in the weights matrix.
Given a multiple sequence alignment between a reference sequence (first sequence in alignment) and a list of structures (remaining sequences in alignment with an attached view to the structure), this set of functions can be used to analyze differences between the structures.
Example:
# SETUP: aln is multiple sequence alignment, where first sequence is the
# reference sequence and all others have a structure attached
# clip alignment to only have parts with at least 3 sequences (incl. ref.)
# > aln will be cut and clip_start is 1st column of aln that was kept
clip_start = seq.alg.ClipAlignment(aln, 3)
# get variance measure and distance to mean for each residue pair
d_map = seq.alg.CreateDistanceMap(aln)
var_map = seq.alg.CreateVarianceMap(d_map)
dist_to_mean = seq.alg.CreateDist2Mean(d_map)
# report min. and max. variances
print("MINMAX:", var_map.Min(), "", var_map.Max())
# get data and jsonstrings for further processing
var_map_data = var_map.GetData()
var_map_json = var_map.GetJsonString()
dist_to_mean_data = dist_to_mean.GetData()
dist_to_mean_json = dist_to_mean.GetJsonString()
ClipAlignment
(aln, n_seq_thresh=2, set_offset=true, remove_empty=true)¶Clips alignment so that first and last column have at least the desired number of structures.
Parameters: 


Returns:  Starting column (0indexed), where cut region starts (w.r.t. original aln). 1, if there is no region in the alignment with at least the desired number of structures. 
Return type: 

CreateDistanceMap
(aln)¶Create distance map from a multiple sequence alignment.
The algorithm requires that the sequence alignment consists of at least two
sequences. The sequence at index 0 serves as a frame of reference. All the
other sequences must have an attached view and a properly set sequence offset
(see SetSequenceOffset()
).
For each of the attached views, the Calpha distance pairs are extracted and mapped onto the corresponding Calpha distances in the reference sequence.
Parameters:  aln (AlignmentHandle ) – Multiple sequence alignment. 

Returns:  Distance map. 
Return type:  DistanceMap 
Raises:  Exception if aln has less than 2 sequences or any sequence (apart from index 0) is lacking an attached view. 
CreateVarianceMap
(d_map, sigma=25)¶Returns:  Variance measure for each entry in d_map. 

Return type:  
Parameters: 

Raises:  Exception if d_map has no entries. 
CreateDist2Mean
(d_map)¶Returns:  Distances to mean for each structure in d_map. Structures are in the same order as passed when creating d_map. 

Return type:  Dist2Mean 
Parameters:  d_map (DistanceMap ) – Distance map as created with CreateDistanceMap() . 
Raises:  Exception if d_map has no entries. 
CreateMeanlDDTHA
(d_map)¶Returns:  lDDT calculation based on CA carbons of the structures with lddt distance threshold of 15 Angstrom and distance difference thresholds of [0.5, 1.0, 2.0, 4.0]. The reported values for a certain structure are the mean perresidue lDDT values given all other structures as reference. Structures are in the same order as passed when creating d_map. 

Return type:  MeanlDDT 
Parameters:  d_map (DistanceMap ) – Distance map as created with CreateDistanceMap() . 
Raises:  Exception if d_map has no entries. 
Distances
¶Container used by DistanceMap
to store a pair wise distance for each
structure. Each structure is identified by its index in the originally used
alignment (see CreateDistanceMap()
).
GetDataSize
()¶Returns:  Number of pairwise distances. 

Return type:  int 
GetAverage
()¶Returns:  Average of all distances. 

Return type:  float 
Raises:  Exception if there are no distances. 
GetMin
()¶GetMax
()¶Returns:  Minimal/maximal distance. 

Return type:  tuple (distance (float ), index (int )) 
Raises:  Exception if there are no distances. 
GetDataElement
(index)¶Returns:  Element at given index. 

Return type:  tuple (distance (float ), index (int )) 
Parameters:  index (int ) – Index within list of distances (must be < GetDataSize() ). 
Raises:  Exception if there are no distances or index out of bounds. 
GetStdDev
()¶Returns:  Standard deviation of all distances. 

Return type:  float 
Raises:  Exception if there are no distances. 
GetWeightedStdDev
(sigma)¶Returns:  Standard deviation of all distances multiplied by
exp( GetAverage() / (2*sigma) ). 

Return type:  float 
Parameters:  sigma (float ) – Defines weight. 
Raises:  Exception if there are no distances. 
GetNormStdDev
()¶Returns:  Standard deviation of all distances divided by GetAverage() . 

Return type:  float 
Raises:  Exception if there are no distances. 
DistanceMap
¶Container returned by CreateDistanceMap()
.
Essentially a symmetric GetSize()
x GetSize()
matrix containing
up to GetNumStructures()
distances (list stored as Distances
).
Indexing of residues starts at 0 and corresponds to the positions in the
originally used alignment (see CreateDistanceMap()
).
GetDistances
(i_res1, i_res2)¶Returns:  List of distances for given pair of residue indices. 

Return type:  
Parameters: 

GetSize
()¶Returns:  Number of residues in map. 

Return type:  int 
GetNumStructures
()¶Returns:  Number of structures originally used when creating the map
(see CreateDistanceMap() ). 

Return type:  int 
VarianceMap
¶Container returned by CreateVarianceMap()
.
Like DistanceMap
, it is a symmetric GetSize()
x GetSize()
matrix containing variance measures.
Indexing of residues is as in DistanceMap
.
Get
(i_res1, i_res2)¶Returns:  Variance measure for given pair of residue indices. 

Return type: 

Parameters: 

GetSize
()¶Returns:  Number of residues in map. 

Return type:  int 
ExportDat
(file_name)¶ExportCsv
(file_name)¶ExportJson
(file_name)¶Write all variance measures into a file. The possible formats are:
GetJsonString()
)Parameters:  file_name (str ) – Path to file to be created. 

Raises:  Exception if the file cannot be opened for writing. 
GetJsonString
()¶Returns:  A JSON formatted list of GetSize() lists with
GetSize() variances 

Return type:  str 
GetData
()¶Gets all the data in this map at once. Note that this is much faster (10x
speedup observed) than parsing GetJsonString()
or using Get()
on each element.
Returns:  A list of GetSize() lists with GetSize() variances. 

Return type:  list of list of float 
Dist2Mean
¶Container returned by CreateDist2Mean()
.
Stores distances to mean for GetNumResidues()
residues of
GetNumStructures()
structures.
Indexing of residues is as in DistanceMap
.
Indexing of structures goes from 0 to GetNumStructures()
 1 and is in
the same order as the structures in the originally used alignment.
Get
(i_res, i_str)¶Returns:  Distance to mean for given residue and structure indices. 

Return type: 

Parameters: 

GetNumResidues
()¶Returns:  Number of residues. 

Return type:  int 
GetNumStructures
()¶Returns:  Number of structures. 

Return type:  int 
ExportDat
(file_name)¶ExportCsv
(file_name)¶ExportJson
(file_name)¶Write all distance measures into a file. The possible formats are:
GetJsonString()
)Parameters:  file_name (str ) – Path to file to be created. 

Raises:  Exception if the file cannot be opened for writing. 
GetJsonString
()¶Returns:  A JSON formatted list of GetNumResidues() lists with
GetNumStructures() distances. 

Return type:  str 
GetData
()¶Gets all the data in this map at once. Note that this is much faster (10x
speedup observed) than parsing GetJsonString()
or using Get()
on each element.
Returns:  A list of GetNumResidues() lists with
GetNumStructures() distances. 

Return type:  list of list of float 
GetSubData
(num_res_to_avg)¶Gets subset of data in this map by averaging neighboring values for num_res_to_avg residues.
Returns:  A list of ceil(GetNumResidues() /num_res_to_avg) lists with
GetNumStructures() distances. 

Return type:  list of list of float 
MeanlDDT
¶Container returned by CreateMeanlDDTHA()
.
Stores mean lDDT values for GetNumResidues()
residues of
GetNumStructures()
structures.
Has the exact same functionality and behaviour as Dist2Mean
Openstructure implements basic HMMrelated functionality that aims at
calculating an HMMHMM alignment score as described in
Soding, Bioinformatics (2005) 21(7), 95160. This is the score which is
optimized in the Viterbi algorithm of the hhalign tool.
As a prerequisite, OpenStructure also implements adding pseudo counts to
ost.seq.ProfileHandle
in order to avoid zero probabilities for
unobserved transitions/emissions. Given these requirements, all functions
in this section require HMM related data (transition probabilities, neff values,
etc.) to be set, which is the case if you load a file in hhm format.
HMMScore
(profile_0, profile_1, aln, s_0_idx, s_1_idx, match_score_offset=0.03, correl_score_weight=0.1, del_start_penalty_factor=0.6, del_extend_penalty_factor=0.6, ins_start_penalty_factor=0.6, ins_extend_penalty_factor=0.6)¶Scores an HMMHMM alignment given in aln between profile_0 and profile_1. The score is described in Soding, Bioinformatics (2005) 21(7), 95160 and consists of three components:
 sum of column alignment scores of all aligned columns, the match_score_offset is applied to each of those scores
 sum of transition probability scores, the prefactor of those scores can be controlled with penalty factors (del_start_penalty_factor etc.)
 correlation score which rewards conserved columns occuring in clusters, correl_score_weight controls its contribution to the total score
You have to make sure that proper pseudo counts are already assigned before calling this function. You can find a usage example in this documentation. This score is not necessarily consistent with the output generated with hhalign, i.e. you take the hhalign output alignment and directly feed it into this function with the same profiles and expect an equal score. The reason is that by default, hhalign performs a realignment step but the output score actually relates to the initial alignment coming from the Viterbi alignment. To get consistent results, run hhalign with the norealign flag.
Parameters: 


Raises:  Exception if profiles don’t have HMM information assigned or specified sequences in aln don’t match with profile SEQRES. Potentially set sequence offsets are taken into account. 
Example with pseudo count assignment:
from ost import io, seq
prof_query = io.LoadSequenceProfile("query.hhm")
prof_tpl = io.LoadSequenceProfile("tpl.hhm")
aln = io.LoadAlignment("aln.fasta")
# assign pseudo counts to transition probabilities
seq.alg.AddTransitionPseudoCounts(prof_query)
seq.alg.AddTransitionPseudoCounts(prof_tpl)
# hhblits/hhalign 3 assign different pseudo counts to
# query and template. The reason is computational efficiency.
# The more expensive Angermueller et al. pseudo counts
# are assigned to the query.
path_to_crf = "/path/to/hhsuite/data/context_data.crf"
lib = seq.alg.ContextProfileDB.FromCRF(path_to_crf)
seq.alg.AddAAPseudoCounts(prof_query, lib)
# templates are assigned the computationally cheaper pseudo
# counts derived from a Gonnet substitution matrix
seq.alg.AddAAPseudoCounts(prof_tpl)
# assign null model pseudo counts
# this should be done AFTER you assigned pseudo counts to emission
# probabilities as this affects the result
seq.alg.AddNullPseudoCounts(prof_query)
seq.alg.AddNullPseudoCounts(prof_tpl)
print("score:", seq.alg.HMMScore(prof_query, prof_tpl, aln, 0, 1))
AddNullPseudoCounts
(profile)¶Adds pseudo counts to null model in profile as implemented in hhalign. Conceptually we’re mixing the original null model with the frequencies observed in the columns of profile. The weight of the original null model depends on the neff value of profile. This function should be called AFTER you already assigned pseudo counts to the emission probabilities as this affects the result.
Parameters:  profile (ost.seq.ProfileHandle ) – Profile to add pseudo counts 

Raises:  Exception if profile doesn’t have HMM information assigned 
AddTransitionPseudoCounts
(profile, gapb=1.0, gapd=0.15, gape=1.0)¶Adds pseudo counts to the transition probabilities in profile as implemented in hhalign with equivalent parameter naming and default parameterization. The original transition probabilities are mixed with prior probabilities that are controlled by gapd and gape. Priors:
 priorM2I = priorM2D = gapd * 0.0286
 priorM2M = 1.0  priorM2D  priorM2I
 priorI2I = priorD2D = 1.0 * gape / (gape  1.0 + 1.0/0.75)
 priorI2M = priorD2M = 1.0  priorI2I
Transition probabilities of column i starting from a match state are then estimated with pM2X = (neff[i]  1) * pM2X + gape * priorM2X. Starting from an insertion/deletion state we have pI2X = neff_ins[i] * pI2X + gape * priorI2X. In the end, all probabilities are normalized such that (pM2M, pM2I, pM2D) sum up to one, (pI2M, pI2I) sum up to one and (pD2I, pD2D) sum up to one.
Parameters:  profile (ost.seq.ProfileHandle ) – Profile to add pseudo counts 

Raises:  Exception if profile doesn’t have HMM information assigned 
AddAAPseudoCounts
(profile, a=1.0, b=1.5, c=1.0)¶Adds pseudo counts to the emission probabilities in profile by mixing in probabilities from the Gonnet matrix as implemented in hhalign with equivalent parameter naming and default parameterization. We only implement the diversitydependent mode for the mixing factor tau (default in hhalign), which for column i depends on neff[i] , a , b and c .
Parameters:  profile (ost.seq.ProfileHandle ) – Profile to add pseudo counts 

Raises:  Exception if profile doesn’t have HMM information assigned 
ContextProfileDB
¶Database that contains context profiles which will be used to add pseudo counts as described by Angermueller et al., Bioinformatics (2012) 28, 32403247.
FromCRF
(filename)¶Static load function which reads a crf file provided in an hhsuite installation. Default location: “path/to/hhsuite/data/context_data.crf”
Parameters:  filename (str ) – Filename of CRF file 

Save
(filename)¶Saves database in OSTinternal binary format which can be loaded faster than a crf file.
Parameters:  filename (str ) – Filename to save db 

Load
(filename)¶Static load function that loads database in OSTinternal binary format.
Parameters:  filename (str ) – Filename of db 

AddAAPseudoCounts
(profile, db, a=0.9, b=4.0, c=1.0)Adds pseudo counts to the emission probabilities in profile by utilizing context profiles as described in Angermueller et al., Bioinformatics (2012) 28, 32403247. We only implement the diversitydependent mode for the mixing factor tau (default in hhalign), which for column i depends on neff[i] , a , b and c .
Parameters: 


Raises:  Exception if profile doesn’t have HMM information assigned 
Enter search terms or a module, class or function name.
seq
– Sequences and Alignments
seq.alg
– Algorithms for Sequences