obitools/obitools3
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 66 Wiki Activity

Fixed the new alignpaired end to work after ngsfilter with the 9879847

Browse Source 
possible cases
This commit is contained in:
Celine Mercier
2019-02-17 18:32:35 +01:00
parent 4ddd1a1c37
commit e026e9ec83
9 changed files with 151 additions and 83 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
python/obitools3/commands/alignpairedend.pyx
View File

@ -15,16 +15,12 @@ from obitools3.libalign._qsrassemble import QSolexaRightReverseAssemble
from obitools3.libalign._solexapairend import buildConsensus, buildJoinedSequence
from obitools3.dms.obiseq cimport Nuc_Seq
from obitools3.libalign.shifted_ali cimport Kmer_similarity, Ali_shifted
from obitools3.commands.ngsfilter import REVERSE_SEQ_COLUMN_NAME, REVERSE_QUALITY_COLUMN_NAME
import sys
import os
REVERSE_SEQ_COLUMN_NAME = b"REVERSE_SEQUENCE"
REVERSE_QUALITY_COLUMN_NAME = b"REVERSE_QUALITY"
__title__="Aligns paired-ended reads"
@ -106,8 +102,7 @@ def alignmentIterator(entries, aligner):
else:
seqF = Nuc_Seq.new_from_stored(entries[i])
seqR = Nuc_Seq(seqF.id, seqF[REVERSE_SEQ_COLUMN_NAME], quality=seqF[REVERSE_QUALITY_COLUMN_NAME])
seqF.pop(REVERSE_SEQ_COLUMN_NAME)
seqF.pop(REVERSE_QUALITY_COLUMN_NAME)
seqR.index = i
ali = aligner(seqF, seqR)
@ -196,7 +191,7 @@ def run(config):
reverse = entries[1]
aligner = Kmer_similarity(forward, view2=reverse, kmer_size=config['alignpairedend']['kmersize'])
else:
aligner = Kmer_similarity(entries, kmer_size=config['alignpairedend']['kmersize'])
aligner = Kmer_similarity(entries, column2=entries[REVERSE_SEQ_COLUMN_NAME], qual_column2=entries[REVERSE_QUALITY_COLUMN_NAME], kmer_size=config['alignpairedend']['kmersize'])
ba = alignmentIterator(entries, aligner)

2
python/obitools3/dms/capi/kmer_similarity.pxd
View File

@ -32,6 +32,8 @@ cdef extern from "kmer_similarity.h" nogil:
OBIDMS_column_p column2,
index_t idx2,
index_t elt_idx2,
OBIDMS_column_p qual_col1,
OBIDMS_column_p qual_col2,
uint8_t kmer_size,
int32_t* kmer_pos_array,
int32_t* kmer_pos_array_height_p,

3
python/obitools3/dms/obiseq.pxd
View File

@ -5,6 +5,7 @@ from .view.view cimport Line
cdef class Seq(dict) :
cdef int _index
cpdef object clone(self)
cpdef str get_str(self)
cpdef get_symbol_at(self, int pos)
@ -22,6 +23,7 @@ cdef class Nuc_Seq(Seq) :
cdef class Seq_Stored(Line) :
cpdef get_symbol_at(self, int pos)
cpdef get_slice(self, slice slice_to_get)
@ -31,6 +33,7 @@ cdef class Nuc_Seq_Stored(Seq_Stored) :
cdef Nuc_Seq _reverse_complement
cdef object _quality_array
cdef bytes _seq
cpdef set(self, object id, object seq, object definition=*, object quality=*, int offset=*, object tags=*)
cpdef set_quality_int(self, list new_qual)
cpdef set_quality_char(self, object new_qual, int offset=*)

14
python/obitools3/dms/obiseq.pyx
View File

@ -40,6 +40,7 @@ cdef class Seq(dict) :
self.id = id
self.seq = seq
self.definition = definition
self._index = -1
if tags is not None :
for k in tags:
k_b = tobytes(k)
@ -54,6 +55,7 @@ cdef class Seq(dict) :
def new_from_stored(Seq_Stored seq_to_clone) :
cdef Seq new_seq
new_seq = Seq(seq_to_clone.id, seq_to_clone.seq, definition=seq_to_clone.definition, quality=seq_to_clone.quality, tags=seq_to_clone)
new_seq._index = seq_to_clone._index
return new_seq
def __contains__(self, object key):
@ -128,7 +130,16 @@ cdef class Seq(dict) :
if new_definition is not None:
new_definition = tobytes(new_definition)
self[DEFINITION_COLUMN] = new_definition
# sequence index (for reference in a view eventually) property getter and setter
@property
def index(self): # @ReservedAssignment
return self._index
@index.setter
def index(self, int idx): # @DuplicatedSignature
self._index = idx
cdef class Nuc_Seq(Seq) :
@ -160,6 +171,7 @@ cdef class Nuc_Seq(Seq) :
def new_from_stored(Nuc_Seq_Stored seq_to_clone) :
cdef Nuc_Seq new_seq
new_seq = Nuc_Seq(seq_to_clone.id, seq_to_clone.seq, definition=seq_to_clone.definition, quality=seq_to_clone.quality, tags=seq_to_clone)
new_seq._index = seq_to_clone.index
return new_seq
# is_revcomp property getter and setter (boolean indicating whether the sequence was created by reverse complementing another sequence)

19
python/obitools3/libalign/_solexapairend.pyx
View File

@ -13,7 +13,7 @@ from obitools3.dms.capi.obidmscolumn cimport OBIDMS_column_p
from obitools3.dms.view.view cimport View
from obitools3.dms.column.column cimport Column
from obitools3.commands.ngsfilter import REVERSE_SEQ_COLUMN_NAME, REVERSE_QUALITY_COLUMN_NAME
from math import log10
@ -182,7 +182,7 @@ def buildConsensus(ali, seq, ref_tags=None):
col_p = column.pointer()
# doesn't work because uint8_t* are forced into bytes by cython (nothing read/kept beyond 0 values)
#obi_set_qual_int_with_elt_idx_and_col_p_in_view(view_p, col_p, seq.index, 0, ali.consensus_qual, ali.consensus_len)
seq.set(ref_tags.id+b"_CONS", ali.consensus_seq, quality=ali.consensus_qual, tags=ref_tags)
seq.set(ref_tags.id+b"_CONS", ali.consensus_seq, quality=ali.consensus_qual)
seq[b'ali_length'] = ali.consensus_len
seq[b'score_norm']=float(ali.score)/ali.consensus_len
seq[b'shift']=ali.shift
@ -208,7 +208,7 @@ def buildConsensus(ali, seq, ref_tags=None):
quality[i] = min(42, round(-10*log10(p)))
i+=1
seq.set(ali[0].wrapped.id+b"_CONS", sseq, quality=quality, tags=ali[0].wrapped)
seq.set(ali[0].wrapped.id+b"_CONS", sseq, quality=quality)
if hasattr(ali, "counter"):
seq[b'alignement_id']=ali.counter
@ -224,12 +224,18 @@ def buildConsensus(ali, seq, ref_tags=None):
seq[b'ali_length']=len(seq)-ic.seqASingle-ic.seqBSingle
if seq[b'ali_length']>0:
seq[b'score_norm']=float(ali.score)/seq[b'ali_length']
ref_tags = ali[0].wrapped
if hasattr(ali, "direction"):
seq[b'ali_direction']=ali.direction
seq[b'score']=ali.score
seq[b'mode']=b'alignment'
for tag in ref_tags:
if tag != REVERSE_SEQ_COLUMN_NAME and tag != REVERSE_QUALITY_COLUMN_NAME:
seq[tag] = ref_tags[tag]
return seq
@ -241,10 +247,13 @@ def buildJoinedSequence(ali, reverse, seq, forward=None):
s = forward.seq + reverse.seq
quality = forward.quality
quality.extend(reverse.quality)
seq.set(forward.id +b"_PairedEnd", s, definition=forward.definition, quality=quality, tags=forward)
seq.set(forward.id +b"_PairedEnd", s, definition=forward.definition, quality=quality)
seq[b"score"]=ali.score
seq[b"ali_direction"]=ali.direction
seq[b"mode"]=b"joined"
seq[b"pairedend_limit"]=len(forward)
for tag in forward:
if tag != REVERSE_SEQ_COLUMN_NAME and tag != REVERSE_QUALITY_COLUMN_NAME:
seq[tag] = forward[tag]
return seq

2
python/obitools3/libalign/shifted_ali.pxd
View File

@ -24,7 +24,9 @@ cdef class Kmer_similarity:
cdef int32_t shift_count_array_height_a[1]
cdef Obiview_p view1_p
cdef OBIDMS_column_p column1_p
cdef OBIDMS_column_p qual_col1_p
cdef Obiview_p view2_p
cdef OBIDMS_column_p column2_p
cdef OBIDMS_column_p qual_col2_p
cdef bint build_consensus
cpdef free(self)

27
python/obitools3/libalign/shifted_ali.pyx
View File

@ -1,8 +1,7 @@
#cython: language_level=3
from obitools3.dms.capi.obiview cimport NUC_SEQUENCE_COLUMN
from obitools3.dms.capi.obiview cimport NUC_SEQUENCE_COLUMN, QUALITY_COLUMN
from obitools3.dms.obiseq cimport Nuc_Seq_Stored
from obitools3.dms.capi.kmer_similarity cimport kmer_similarity, Obi_ali_p, obi_free_shifted_ali
from obitools3.dms.view.typed_view.view_NUC_SEQS cimport View_NUC_SEQS
from obitools3.dms.column.column cimport Column
@ -52,13 +51,14 @@ cdef class Ali_shifted:
def direction(self):
return self.pointer().direction
cpdef free(self):
cpdef free(self): # TODO allocated memory could be kept
obi_free_shifted_ali(self.pointer())
cdef class Kmer_similarity:
def __init__(self, View_NUC_SEQS view1, Column column1=None, View_NUC_SEQS view2=None, Column column2=None, uint8_t kmer_size=3, build_consensus=True) :
def __init__(self, View_NUC_SEQS view1, Column column1=None, Column qual_column1=None, View_NUC_SEQS view2=None, Column column2=None, Column qual_column2=None, uint8_t kmer_size=3, build_consensus=True) :
cdef Column default_seq_col
cdef Column default_qual_col
if kmer_size < 1 or kmer_size > 3:
raise Exception("Kmer size to compute kmer similarity must be >=1 or <=4")
self.kmer_pos_array_p = NULL
@ -89,13 +89,28 @@ cdef class Kmer_similarity:
raise Exception("Kmer similarity with no sequence column given must be given a NUC_SEQS view")
default_seq_col = view2[NUC_SEQUENCE_COLUMN]
self.column2_p = default_seq_col.pointer()
if qual_column1 is not None:
self.qual_col1_p = qual_column1.pointer()
else:
if type(view1) != View_NUC_SEQS:
raise Exception("Kmer similarity with no quality column given must be given a NUC_SEQS view")
default_qual_col = view1[QUALITY_COLUMN]
self.qual_col1_p = default_qual_col.pointer()
if qual_column2 is not None:
self.qual_col2_p = qual_column2.pointer()
else:
if type(view2) != View_NUC_SEQS:
raise Exception("Kmer similarity with no quality column given must be given a NUC_SEQS view")
default_qual_col = view2[QUALITY_COLUMN]
self.qual_col2_p = default_qual_col.pointer()
def __call__(self, Nuc_Seq_Stored seq1, Nuc_Seq_Stored seq2):
def __call__(self, object seq1, object seq2):
cdef Obi_ali_p ali_p
cdef Ali_shifted ali
ali_p = kmer_similarity(self.view1_p, self.column1_p, seq1.index, 0, \
self.view2_p, self.column2_p, seq2.index, 0, \
self.qual_col1_p, self.qual_col2_p, \
self.kmer_size, \
self.kmer_pos_array_p, self.kmer_pos_array_height_a, \
self.shift_array_p, self.shift_array_height_a, \