Difference between revisions of "Collection of scripts"

From Bioinformatikpedia
(ss_score.py)
 
(11 intermediate revisions by the same user not shown)
Line 1: Line 1:
A completely unordered, yet to become truly usefull collection of scripts [[Phenylketonuria|we]] used.
+
A completely unordered, yet to become truly useful collection of scripts [[Phenylketonuria|we]] used.
   
  +
==Sequence Alignments==
==PSIBlasttrainer.pl==
 
  +
===PSIBlasttrainer.pl===
  +
<source lang="perl">
 
#! /usr/bin/perl
 
#! /usr/bin/perl
 
use strict;
 
use strict;
Line 54: Line 56:
 
 
 
close TIME;
 
close TIME;
  +
</source>
   
  +
===Sequence Alignment Statistics===
 
  +
<source lang="python">
==Sequence Alignments==
 
 
#!/usr/bin/env python
 
#!/usr/bin/env python
 
#simple sript to count gaps and conserved columns in multiple alignment files
 
#simple sript to count gaps and conserved columns in multiple alignment files
Line 115: Line 118:
 
print "Only .aln or .fasta files, sorry!"
 
print "Only .aln or .fasta files, sorry!"
   
  +
</source>
   
==HHextract.py==
+
===HHextract.py===
  +
<source lang="python">
 
 
#!/usr/bin/env python
 
#!/usr/bin/env python
 
#this simply takes output of HHblits, extracts the IDs of cluster members and adds a clusteridentifier
 
#this simply takes output of HHblits, extracts the IDs of cluster members and adds a clusteridentifier
Line 160: Line 164:
 
infile.close()
 
infile.close()
 
outfile.close()
 
outfile.close()
  +
  +
</source>
  +
  +
===getEvalues.pl===
  +
<source lang="perl">
  +
#!/usr/bin/perl
  +
use strict;
  +
use DirHandle;
  +
  +
my $dir = $ARGV[0];
  +
my $blast = $ARGV[1];
  +
  +
sub plainfiles {
  +
my $dir = shift;
  +
my $dh = DirHandle->new($dir) or die "can't opendir $dir: $!";
  +
return sort # sort pathnames
  +
grep { -f } # choose only "plain" files
  +
map { "$dir/$_" } # create full paths
  +
grep { !/^\./ } # filter out dot files
  +
grep { !/\.dat$/}
  +
grep { !/\.pl$/}
  +
grep { !/\.py$/}
  +
grep { !/\.png$/}
  +
$dh->
  +
read()
  +
; # read all entries
  +
}
  +
  +
my @files = plainfiles($dir);
  +
  +
if($blast eq "blast"){
  +
for(@files){
  +
open(READER ,"<",$_);
  +
open(WRITER ,">",$_.".dat");
  +
my $eval;
  +
while(<READER>){
  +
if(/Expect = (.*?),/){
  +
#print "buuuuu";
  +
if($1 == 0){
  +
$eval = "-INF";
  +
}else{
  +
$eval = log($1)/log(10);
  +
}
  +
#Identities = 314/440 (71%)
  +
}elsif(/Identities = .*?\/.*? \((.*?)\%\),/){
  +
#print "tralaöa";
  +
print WRITER "$eval 0\.$1\n";
  +
}elsif(/Results of round/){
  +
close WRITER;
  +
open(WRITER ,">",$_.".dat");
  +
}
  +
}
  +
close READER;
  +
close WRITER;
  +
}
  +
}else{
  +
for(@files){
  +
open(READER ,"<",$_);
  +
open(WRITER ,">",$_.".dat");
  +
while(<READER>){
  +
  +
my @fields = split(" ",$_);
  +
my $eval;
  +
if($fields[2] == 0){
  +
$eval = "-INF";
  +
}else{
  +
$eval = log($fields[2])/log(10);
  +
}
  +
my $ident =$fields[3];
  +
$ident =~ s/\%//;
  +
$ident = "0."+$ident;
  +
print WRITER "$eval $ident\n";
  +
  +
}
  +
close READER;
  +
close WRITER;
  +
}
  +
}
  +
  +
</source>
  +
  +
==Sequence based predictions==
  +
===ss_score.py===
  +
<source lang="python">
  +
#!/usr/bin/env python
  +
#calculates Q3 and SOV for two (already aligned) assignments of structure
  +
#in files containing only the structure (no comment line, no header, mayyybe newlines..)
  +
#usage: python ss_score.py <dssp_file> <prediction_file>
  +
  +
import sys
  +
  +
def main():
  +
#read files, map DSSP characters to EHL and C to L
  +
#strip of newlines..
  +
dssp = open(sys.argv[1]).read().strip().replace("S","L").replace("B","L").replace("I","L").replace("G","H").replace("T","L").replace("\n","")
  +
pred_tmp = open(sys.argv[2]).read().strip().replace("C","L").replace("\n","")
  +
  +
#if DSSP has only '-', ignore in prediction as well
  +
pred = ""
  +
for i in range(len(dssp)):
  +
if dssp[i] == "-":
  +
pred+= "-"
  +
else:
  +
pred+=pred_tmp[i]
  +
  +
q_values = q3(dssp,pred)
  +
sov_values = sov(dssp,pred)
  +
print "Q3 =", q_values[0],"\nQ_E =", q_values[1],"\nQ_H =", q_values[2],"\nQ_L =", q_values[3]
  +
print "SOV =", sov_values[0],"\nSOV_E =", sov_values[1],"\nSOV_H =", sov_values[2],"\nSOV_L =", sov_values[3]
  +
  +
def q3(dssp,pred):
  +
residues_E = 0
  +
residues_H = 0
  +
residues_L = 0
  +
match_E = 0
  +
match_H = 0
  +
match_L = 0
  +
residues = {'E':residues_E,'H':residues_H,'L':residues_L}
  +
match = {'E':match_E,'H':match_H,'L':match_L}
  +
for i in range(len(dssp)):
  +
if dssp[i] != "-":
  +
residues[dssp[i]]+=1
  +
if dssp[i] == pred[i]:
  +
match[dssp[i]]+=1
  +
matches = sum(match.values())
  +
resid = sum(residues.values())
  +
return (matches*100.0/resid,match['E']*100.0/residues['E'],match['H']*100.0/residues['H'],match['L']*100.0/residues['L'])
  +
  +
def sov(dssp,pred):
  +
sov = 0
  +
sov_E = 0
  +
sov_H = 0
  +
sov_L = 0
  +
segment_sum = 0
  +
for typ in ["E","H","L"]:
  +
stretch_A = []
  +
stretch_B = []
  +
  +
started = False
  +
start = 0
  +
  +
sov_x = 0
  +
segment_x = 0
  +
  +
for pairs in [(stretch_A, dssp),(stretch_B, pred)]:
  +
seq = pairs[1]
  +
sets = pairs[0]
  +
for i in range(len(seq)):
  +
if not started and seq[i] == typ:
  +
started = True
  +
start = i
  +
  +
if started and seq[i] != typ:
  +
started = False
  +
#first position and first position after
  +
sets.append((start,i))
  +
start = 0
  +
for segment in stretch_A:
  +
segment_x+=segment[1]-segment[0]
  +
overlapping_pairs = []
  +
a = 0
  +
b = 0
  +
while a < len(stretch_A) and b < len(stretch_B):
  +
a_start = stretch_A[a][0]
  +
a_end = stretch_A[a][1]
  +
  +
b_start = stretch_B[b][0]
  +
b_end = stretch_B[b][1]
  +
if a_start <= b_start:
  +
if b_start <= a_end:
  +
overlapping_pairs.append((stretch_A[a],stretch_B[b]))
  +
b+=1
  +
else:
  +
a+=1
  +
else:
  +
if a_start <= b_end:
  +
overlapping_pairs.append((stretch_A[a],stretch_B[b]))
  +
a+=1
  +
else:
  +
b+=1
  +
  +
segment_sum+=segment_x
  +
  +
for pair in overlapping_pairs:
  +
sov_x+= (pair[0][1]-pair[0][0])*(minov(pair)+delta(pair))/maxov(pair)
  +
sov+=sov_x
  +
if typ == 'E':
  +
sov_E = sov_x*100/segment_x
  +
elif typ == 'H':
  +
sov_H = sov_x*100/segment_x
  +
elif typ =='L':
  +
sov_L = sov_x*100/segment_x
  +
return (sov*100/segment_sum,sov_E,sov_H, sov_L)
  +
  +
def minov(pair):
  +
a = pair[0]
  +
b = pair[1]
  +
return float(min(a[1],b[1]) - max(a[0],b[0]))
  +
  +
def maxov(pair):
  +
a = pair[0]
  +
b = pair[1]
  +
return float(max([a[1],b[1]]) - min([a[0],b[0]]))
  +
  +
def delta(pair):
  +
return float(min([maxov(pair)-minov(pair),minov(pair),(pair[0][1]-pair[0][0])/2,(pair[1][1]-pair[1][0])/2]))
  +
  +
main()
  +
</source>
  +
  +
[[Category: Phenylketonuria 2012]]

Latest revision as of 12:01, 29 August 2012

A completely unordered, yet to become truly useful collection of scripts we used.

Sequence Alignments

PSIBlasttrainer.pl

<source lang="perl">

#! /usr/bin/perl
use strict;
use Getopt::Long;

my $input;
my $output;
my $database;
my $maxIt;
my $before;
my $after;

my $result = GetOptions ("i=s" => \$input,"o=s" => \$output,"d=s" => \$database,"j=i" => \$maxIt);
open(TIME ,">","time.txt");
print TIME "\# time used for psiblast\n";

#run psiblast each time and save the time;
for(my $i =1;$i<=$maxIt;$i++){
	print "iteration: $i\n";
	$before = time();
	open STATUS, "-|","time blastpgp -j $i -d $database -i $input -o $output"."_".$i."_it" 
		or die "can't fork: $!";
	while(<STATUS>){
		print TIME $_;
	}	
	$after = time();
	print TIME ($after-$before)."\n";
	close STATUS;
}
print TIME ">";

for(my $i=1;$i<$maxIt;$i++){
	my $eval = 1/(10**$i);
	print "iteration: $i\n";
	$before = time();
	system("time blastpgp -j 1 -d $database -i $input -o $output" . "_eVal_" . "$eval -e $eval");
	$after = time();
	print TIME ($after-$before)."\n";
	close STATUS;
} 

print TIME ">\n"
for(my $i=1;$i<$maxIt;$i++){
	my $eval = 1/(10**$i);
	print "iteration: $i\n";
	$before = time();
	system("time blastpgp -j 1 -d $database -i $input -o $output" . "_integr_" . "$eval -h $eval");
	$after = time();
	print TIME ($after-$before)."\n";
	close STATUS;
}
 
close TIME;

</source>

Sequence Alignment Statistics

<source lang="python">

#!/usr/bin/env python
#simple sript to count gaps and conserved columns in multiple alignment files
import sys
from collections import defaultdict

def default_string():
    return ""

def aln_stats(filename):
    fh = open(filename,'r')
    sequences = defaultdict(default_string)
    for line in fh:
        if line[:2].strip() and not line.startswith("CLUSTAL"):
            seq = line.split()
            sequences[seq[0]]+="".join(seq[1:])
    fh.close()

    for key in sequences:
        print key, sequences[key].count('-')
    print "conserved:",conserved(sequences)
    
def fasta_stats(filename):
    fh = open(filename,'r')
    gap_count = 0
    seq_id = ""
    sequences = defaultdict(default_string)
    for line in fh:
        if line.startswith(">"):
            if gap_count:
                print seq_id, gap_count
            gap_count = 0
            seq_id = line.split()[0][1:]
        else:
            gap_count+=line.count('-')
            sequences[seq_id]+=line.strip()

    print seq_id, gap_count,"\nconserved:",conserved(sequences)
    fh.close() 

def conserved(sequence_dict):
    conserved_count = 0
    for i in range(len(sequence_dict.values()[0])):
        col = set([])
        for seq in sequence_dict.values():
            col.add(seq[i])
        if len(col)== 1:
            conserved_count+=1
    return conserved_count 

if "--help" in sys.argv or len(sys.argv)<2:
    print "usage: python %s <filename>"%sys.argv[0]
elif sys.argv[1].endswith(".aln"):
    aln_stats(sys.argv[1])
elif sys.argv[1].endswith(".fasta"):
    fasta_stats(sys.argv[1])
else:
   print "Only .aln or .fasta files, sorry!"

</source>

HHextract.py

<source lang="python">

#!/usr/bin/env python
#this simply takes output of HHblits, extracts the IDs of cluster members and adds a clusteridentifier
#output:
#ID1 CL1 E-Value Identities
#ID2 CL1 E-Value Identities
#ID3 CL1 E-Value Identities
#ID4 CL2 E-Value Identities
#ID4 CL2 E-Value Identities
#..
#
#usage: python hhextract.py <infile> <outfile>
 
import sys, re
try:
    infile = open(sys.argv[1])
except IOError:
    print "Couldn't open input file"
    exit(-1)
try:
    outfile= open(sys.argv[2],'w')
except Exception:
    print "Couldn't open output file"
    exit(-1)
 
for line in infile:
    if line.startswith('>'):
        splits = line.split('|')
        to_find = int(splits[2])
        #check number of clustermembers, skip and print clusterID if confusion
        #if len(splits)!= (4+to_find):
        #   print line[:15]
        #   print to_find, splits
        #   exit(-1)
        line = infile.next()
        splits2 = line.split()
        for member in splits[-to_find:]:
            memberID = member.split()[0].strip()
            if memberID.endswith('.'):
                memberID = memberID[:-1]
            outfile.write("%s %s %s %s\n"%(memberID, splits[1], splits2[1].split('=')[1], splits2[4].split('=')[1]))
infile.close()
outfile.close()

</source>

getEvalues.pl

<source lang="perl">

#!/usr/bin/perl
use strict;
use DirHandle;

my $dir = $ARGV[0];
my $blast = $ARGV[1];

sub plainfiles {
   my $dir = shift;
   my $dh = DirHandle->new($dir)   or die "can't opendir $dir: $!";
   return sort                     # sort pathnames
         grep {    -f     }       # choose only "plain" files
         map  { "$dir/$_" }       # create full paths
         grep {  !/^\./   }       # filter out dot files
	  grep {  !/\.dat$/}
	  grep {  !/\.pl$/}
	  grep {  !/\.py$/}
	  grep {  !/\.png$/}
          $dh->
read()
;             # read all entries
}

my @files =  plainfiles($dir); 

if($blast eq "blast"){
  for(@files){
    open(READER ,"<",$_);
    open(WRITER ,">",$_.".dat");
    my $eval;
    while(<READER>){
      if(/Expect = (.*?),/){
	 #print "buuuuu";
	 if($1 == 0){
		$eval = "-INF";
	}else{
	 $eval = log($1)/log(10);
	}
	#Identities = 314/440 (71%)
      }elsif(/Identities = .*?\/.*? \((.*?)\%\),/){
	#print "tralaöa";
	print WRITER "$eval 0\.$1\n";
      }elsif(/Results of round/){
	close WRITER; 
	open(WRITER ,">",$_.".dat");
      }
    }
    close READER;
    close WRITER;
  }
}else{
  for(@files){
    open(READER ,"<",$_);
    open(WRITER ,">",$_.".dat");
    while(<READER>){
     
	my @fields = split(" ",$_);
	my $eval;
 	if($fields[2] == 0){
		$eval = "-INF";
	}else{
	 $eval = log($fields[2])/log(10);
	}
	my $ident =$fields[3]; 
	$ident =~ s/\%//;
	$ident = "0."+$ident;  
	print WRITER "$eval $ident\n";
       
    }
    close READER;
    close WRITER;
  }
}

</source>

Sequence based predictions

ss_score.py

<source lang="python">

#!/usr/bin/env python
#calculates Q3 and SOV for two (already aligned) assignments of structure
#in files containing only the structure (no comment line, no header, mayyybe newlines..)
#usage: python ss_score.py <dssp_file> <prediction_file>

import sys

def main():
    #read files, map DSSP characters to EHL and C to L
    #strip of newlines..
    dssp = open(sys.argv[1]).read().strip().replace("S","L").replace("B","L").replace("I","L").replace("G","H").replace("T","L").replace("\n","")
    pred_tmp = open(sys.argv[2]).read().strip().replace("C","L").replace("\n","")

    #if DSSP has only '-', ignore in prediction as well
    pred = ""
    for i in range(len(dssp)):
        if dssp[i] == "-":
            pred+= "-"
        else:
            pred+=pred_tmp[i]

    q_values = q3(dssp,pred)
    sov_values = sov(dssp,pred)
    print "Q3 =", q_values[0],"\nQ_E =", q_values[1],"\nQ_H =", q_values[2],"\nQ_L =", q_values[3]
    print "SOV =", sov_values[0],"\nSOV_E =", sov_values[1],"\nSOV_H =", sov_values[2],"\nSOV_L =", sov_values[3]

def q3(dssp,pred):
    residues_E = 0
    residues_H = 0
    residues_L = 0
    match_E = 0
    match_H = 0
    match_L = 0
    residues = {'E':residues_E,'H':residues_H,'L':residues_L}
    match = {'E':match_E,'H':match_H,'L':match_L}
    for i in range(len(dssp)):
        if dssp[i] != "-":
            residues[dssp[i]]+=1
            if dssp[i] == pred[i]:
                match[dssp[i]]+=1
    matches = sum(match.values())
    resid = sum(residues.values())
    return (matches*100.0/resid,match['E']*100.0/residues['E'],match['H']*100.0/residues['H'],match['L']*100.0/residues['L'])

def sov(dssp,pred):
    sov = 0
    sov_E = 0
    sov_H = 0
    sov_L = 0
    segment_sum = 0
    for typ in ["E","H","L"]:
        stretch_A = []
        stretch_B = []

        started = False
        start = 0
        
        sov_x = 0
        segment_x = 0
        
        for pairs in [(stretch_A, dssp),(stretch_B, pred)]:
            seq = pairs[1]
            sets = pairs[0]
            for i in range(len(seq)):
                if not started and seq[i] == typ:
                    started = True
                    start = i

                if started and seq[i] != typ:
                    started = False
                    #first position and first position after
                    sets.append((start,i))
            start = 0
        for segment in stretch_A:
            segment_x+=segment[1]-segment[0]
        overlapping_pairs = []
        a = 0
        b = 0
        while a < len(stretch_A) and b < len(stretch_B):
            a_start = stretch_A[a][0]
            a_end = stretch_A[a][1]

            b_start = stretch_B[b][0]
            b_end = stretch_B[b][1]
            if a_start <= b_start:
                if b_start <= a_end:
                    overlapping_pairs.append((stretch_A[a],stretch_B[b]))
                    b+=1
                else:
                    a+=1 
            else:
                if a_start <= b_end:
                    overlapping_pairs.append((stretch_A[a],stretch_B[b]))
                    a+=1
                else:
                    b+=1

        segment_sum+=segment_x

        for pair in overlapping_pairs:
            sov_x+= (pair[0][1]-pair[0][0])*(minov(pair)+delta(pair))/maxov(pair)
        sov+=sov_x
        if typ == 'E':
            sov_E = sov_x*100/segment_x
        elif typ == 'H':
            sov_H = sov_x*100/segment_x
        elif typ =='L':
            sov_L = sov_x*100/segment_x
    return (sov*100/segment_sum,sov_E,sov_H, sov_L)

def minov(pair):
    a = pair[0]
    b = pair[1]
    return float(min(a[1],b[1]) - max(a[0],b[0]))

def maxov(pair):
    a = pair[0]
    b = pair[1]
    return float(max([a[1],b[1]]) - min([a[0],b[0]]))

def delta(pair):
    return float(min([maxov(pair)-minov(pair),minov(pair),(pair[0][1]-pair[0][0])/2,(pair[1][1]-pair[1][0])/2])) 

main()

</source>