Class: PluginUserContaminants

Inherits:
Plugin
  • Object
show all
Defined in:
lib/seqtrimnext/plugins/plugin_user_contaminants.rb

Constant Summary collapse

MAX_TARGETS_SEQS =

MAXIMUM NUMBER OF DIFFERENT ALIGNED SEQUENCES TO KEEP FROM BLAST DATABASE

4

Instance Attribute Summary

Attributes inherited from Plugin

#stats

Class Method Summary collapse

Instance Method Summary collapse

Methods inherited from Plugin

#add_stats, #add_text_stats, auto_setup, check_param, #execute, get_graph_filename, get_graph_title, graph_ignored?, ignored_graphs, #initialize, #merge_hits, #overlapX?, plot_setup, valid_graphs

Constructor Details

This class inherits a constructor from Plugin

Class Method Details

.check_params(params) ⇒ Object

Returns an array with the errors due to parameters are missing



127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
 
# File 'lib/seqtrimnext/plugins/plugin_user_contaminants.rb', line 127

def self.check_params(params)
  errors=[]


  comment='Blast E-value used as cut-off when searching for contaminations'
  default_value = 1e-10
  params.check_param(errors,'blast_evalue_user_contaminant','Float',default_value,comment)

  comment='Minimum required identity (%) for a reliable user contaminant match'
  default_value = 85
  params.check_param(errors,'blast_percent_user_contaminant','Integer',default_value,comment)

  comment='Minimum hit size (nt) for considering for user contaminant'
  default_value = 30 # era 40
  params.check_param(errors,'min_user_contaminant_size','Integer',default_value,comment)

  comment='Path for user contaminant database'
  default_value = "" #File.join($FORMATTED_DB_PATH,'user_contaminant.fasta')
  params.check_param(errors,'user_contaminant_db','DB',default_value,comment)

  return errors
end

Instance Method Details

#can_execute? ⇒ Boolean

Returns:

  • (Boolean) 


36
37
38
 
# File 'lib/seqtrimnext/plugins/plugin_user_contaminants.rb', line 36

def can_execute?
  return !@params.get_param('user_contaminant_db').empty?
end

#do_blasts(seqs) ⇒ Object 



41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
 
# File 'lib/seqtrimnext/plugins/plugin_user_contaminants.rb', line 41

def do_blasts(seqs)

  # TODO - Culling limit = 2 porque el blast falla con este comando cuando se le pasa cl=1 y dust=no
  # y una secuencia de baja complejidad como entrada

  blast = BatchBlast.new("-db #{@params.get_param('user_contaminant_db')}",'blastn'," -task blastn -evalue #{@params.get_param('blast_evalue_user_contaminant')} -perc_identity #{@params.get_param('blast_percent_user_contaminant')} -culling_limit 1")  #get classify -max_target_seqs #{MAX_TARGETS_SEQS}

  $LOG.debug('BLAST:'+blast.get_blast_cmd(:xml))

  fastas=[]

  seqs.each do |seq|
    fastas.push ">"+seq.seq_name
    fastas.push seq.seq_fasta
  end


  blast_table_results = blast.do_blast(fastas,:xml)

  return blast_table_results
end

#exec_seq(seq, blast_query) ⇒ Object 



64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
 
# File 'lib/seqtrimnext/plugins/plugin_user_contaminants.rb', line 64

def exec_seq(seq,blast_query)
  if blast_query.query_id != seq.seq_name
    # raise "Blast and seq names does not match, blast:#{blast_query.query_id} sn:#{seq.seq_name}"
  end

  $LOG.debug "[#{self.class.to_s}, seq: #{seq.seq_name}]: looking for classify into the sequence"

  type = "ActionUserContaminant"

  classify={}
  contaminants=[]

  
  merge_hits(blast_query.hits,contaminants,nil,false)

  begin
    contaminants2=contaminants
    contaminants = []                            #second round to save contaminants without overlap
    merge_hits(contaminants2,contaminants,nil,false)
  end until (contaminants2.count == contaminants.count)
  
  contaminants.sort {|c1,c2| (c1.q_end - c1.q_beg + 1)<=>(c2.q_end - c2.q_beg + 1)}

  # classify=sum_hits_by_id(contaminants.hits)

  actions=[]
  # classify_size=0

  min_cont_size=@params.get_param('min_user_contaminant_size').to_i
  
  # biggest_classify = contaminants.sort {|c1,c2| c1[1]<=>c2[1]}
  
  if !contaminants.empty?

    # definition,classify_size = biggest_classify.last
    
    biggest_contaminant=contaminants.last
    hit_size=(biggest_contaminant.q_end - biggest_contaminant.q_beg + 1)
    
    a = seq.new_action(biggest_contaminant.q_beg,biggest_contaminant.q_end,type) # adds the correspondent action to the sequence

    a.message = biggest_contaminant.definition
    
    seq.add_comment("Contaminated: #{biggest_contaminant.definition}")
    
    a.tag_id = biggest_contaminant.definition.gsub(' ','_')

    # a.found_definition = c.definition    # save the classify definitions, each separately
    
    #save to this file
    seq.add_file_tag(0, 'with_user_contaminant', :both, 10)
    
    actions.push a
    
    add_stats('user_contaminant_size',hit_size)
    add_stats('user_contaminant_ids',biggest_contaminant.definition)

    seq.add_actions(actions)
  end

end

#near_to_extrem(c, seq, min_cont_size) ⇒ Object 



17
18
19
20
 
# File 'lib/seqtrimnext/plugins/plugin_user_contaminants.rb', line 17

def near_to_extrem(c,seq,min_cont_size)
  max_to_extreme=(min_cont_size/2).to_i
  return ((c.q_beg-max_to_extreme<0) || (( c.q_end+max_to_extreme)>=seq.seq_fasta.size-1) ) #return if vector is very near to the extremes of insert)
end

#sum_hits_by_id(hits) ⇒ Object 



22
23
24
25
26
27
28
29
30
31
32
33
34
 
# File 'lib/seqtrimnext/plugins/plugin_user_contaminants.rb', line 22

def sum_hits_by_id(hits)
  res={}
  
  hits.each do |c|
    hit_size=c.q_end - c.q_beg + 1
    
    res[c.definition] = (res[c.definition]||0)+hit_size
    
  end
  
  puts res.to_json
  return res
end