|
8
|
1 #!/usr/bin/perl -w
|
|
|
2
|
|
|
3 #Margaret Antonio 16.01.13
|
|
|
4
|
|
|
5 #DESCRIPTION: Takes two aggregate.pl outputs and compares them using mean difference, pval for each
|
|
|
6 #gene. Can compare, for example, 19F in glucose and TIGR4 in glucose.
|
|
|
7 #DIFFERENT GENOMES (ie. diff. strains).
|
|
|
8 #Requires CONVERSION FILE
|
|
|
9
|
|
|
10 #USAGE: perl compStrains.pl -c <conversion.csv> <options>
|
|
|
11 #[<aggregateFile1.csv aggregateFile2.csv> OR -indir <indir/>]
|
|
|
12
|
|
|
13 use Data::Dumper;
|
|
|
14 use strict;
|
|
|
15 use Getopt::Long;
|
|
|
16 use warnings;
|
|
|
17 use File::Path;
|
|
|
18 use File::Basename;
|
|
|
19 use Statistics::Distributions;
|
|
|
20
|
|
|
21 #ASSIGN INPUTS TO VARIABLES USING FLAGS
|
|
|
22 our ($indir,$h,$out,$sortkey,$round,$l,$cfile);
|
|
|
23 GetOptions(
|
|
|
24 'd:s' => \$indir,
|
|
|
25 'h' => \$h,
|
|
|
26 'o:s' =>\$out,
|
|
|
27 's:i' => \$sortkey,
|
|
|
28 'r:i'=> \$round,
|
|
|
29 'l1:s'=> \$l1,
|
|
|
30 'l2:s'=> \$l2,
|
|
|
31 'c:s'=> \$cfile,
|
|
|
32 );
|
|
|
33
|
|
|
34 sub print_usage() {
|
|
|
35 print "\n";
|
|
|
36 print "\n##################################################################\n";
|
|
|
37 print "compStrains.pl: compare genes from a tn-seq experiment\n";
|
|
|
38 print "\tfor two DIFFERENT strains/genomes using aggregate files\n";
|
|
|
39
|
|
|
40 print "\nDESCRIPTION: Takes two aggregate.pl outputs and compares them by\n";
|
|
|
41 print "calculating the difference in mean fitness.\n";
|
|
|
42
|
|
|
43 print "Example: two strains tested under same condition.\n";
|
|
|
44 print "Note: For same strains (genomes), use compGenes.pl\n";
|
|
|
45
|
|
|
46 print "\nUSAGE:\n";
|
|
|
47 print "perl compStrains.pl -c conversion.csv -d inputs/\n";
|
|
|
48
|
|
|
49 print "\nREQUIRED:\n";
|
|
|
50 print " -d\tDirectory containing all input files (files from\n";
|
|
|
51 print " \taggregate fitness script)\n";
|
|
|
52 print " \tOR\n";
|
|
|
53 print " \tIn the command line (without a flag), input the name(s) of\n";
|
|
|
54 print " \ttwo files containing aggregate gene fitness values. \n\n";
|
|
|
55 print " -c\tConversion file: two columns with homologs for both organisms\n";
|
|
|
56
|
|
|
57 print "\nOPTIONAL:\n";
|
|
|
58 print " -h\tPrints usage and exits program\n";
|
|
|
59 print " -o\tOutput file for comparison data. Default: label1label2.csv\n";
|
|
|
60 print " -s\tSort output by this index of the file (indices begin at 0).\n";
|
|
|
61 print " \tDefault: by mean\n";
|
|
|
62 print " -r\tRound final output numbers to this number of decimals\n";
|
|
|
63 print " -l\tLabels for input files. Default: filenames\n";
|
|
|
64 print " \tTwo strings, comma separated (i.e. -l expt1,expt2).\n";
|
|
|
65 print " \tOrder should match file order.\n";
|
|
|
66 print " \n~~~~Always check that file paths are correctly specified~~~~\n";
|
|
|
67 print "\n##################################################################\n";
|
|
|
68 }
|
|
|
69 if ($h){
|
|
|
70 print_usage();
|
|
|
71 exit;
|
|
|
72 }
|
|
|
73 if (!$indir and (scalar @ARGV==0)){
|
|
|
74 print "\nERROR: Please correctly specify input files or directory\n";
|
|
|
75 print_usage();
|
|
|
76 print "\n";
|
|
|
77 exit;
|
|
|
78 }
|
|
|
79 if (!$cfile){
|
|
|
80 print "\nERROR: Please correctly specify the required conversion file\n";
|
|
|
81 print_usage();
|
|
|
82 print "\n";
|
|
|
83 exit;
|
|
|
84 }
|
|
|
85
|
|
|
86 #THE @files ARRAY WILL CONTAIN INPUT FILE NAMES, EXTRACTED FROM A DIRECTORY (-indir) OR ARGV
|
|
|
87 my @files;
|
|
|
88 if ($indir){
|
|
|
89 my $directory="$indir";
|
|
|
90 opendir(DIR, $directory) or (print "Couldn't open $directory: $!\n" and print_usage() and exit);
|
|
|
91 my @direct= readdir DIR;
|
|
|
92 my $tail=".csv";
|
|
|
93 foreach (@direct){
|
|
|
94 if (index($_, $tail) != -1){
|
|
|
95 $_=$indir.$_;
|
|
|
96 push (@files,$_);
|
|
|
97 }
|
|
|
98 }
|
|
|
99 closedir DIR;
|
|
|
100 }
|
|
|
101 else{
|
|
|
102 @files=@ARGV;
|
|
|
103 }
|
|
|
104
|
|
|
105 #GET LABELS: USE (-l) OR USE FILNEAMES AS LABELS FOR COLUMNS IN OUTPUT FILE
|
|
|
106
|
|
|
107
|
|
|
108 my @labels = ($l1,$l2);
|
|
|
109 #if ($l){
|
|
|
110 # @labels=split(',',$l);
|
|
|
111 #}
|
|
|
112 #else{
|
|
|
113 # foreach (@files){
|
|
|
114 # my $filename=basename($_);
|
|
|
115 # my @temp=split('\\.',$filename);
|
|
|
116 # my $colName=$temp[0];
|
|
|
117 # push (@labels,$colName);
|
|
|
118 # }
|
|
|
119 #}
|
|
|
120
|
|
|
121 #CHECK IF REQ. VARIABLES WERE DEFINED USING FLAGS. IF NOT THEN USE DEFAULT VALUES
|
|
|
122
|
|
|
123 if (!$out) {$out="comp.".$labels[0].$labels[1].".csv"}
|
|
|
124 if (!$round){$round='%.4f'}
|
|
|
125
|
|
|
126 #OPEN INPUTTED AGGREGATE GENE FILES AND STORE THEIR CONTENTS INTO TWO HASHES
|
|
|
127 #FILE1 GOES INTO HASH %ONE AND FILE2 GOES INTO HASH %TWO.
|
|
|
128
|
|
|
129 #FILE1 OPENING ---> %one WHERE KEY:VALUE IS GENE_ID:(GENE_ID,INSERTIONS,MEAN,ETC.)
|
|
|
130 my @header;
|
|
|
131 my %one;
|
|
|
132
|
|
|
133 open (F1,'<',$files[0]);
|
|
|
134
|
|
|
135 #STORE COLUMN NAMES (FIRST LINE OF FILE1) FOR HEADER AND APPEND LABELS
|
|
|
136 my $head=<F1>; #the header in the file
|
|
|
137 my @cols=split(',',$head);
|
|
|
138 @cols=@cols[0,1,2,3,4,5,6]; #get rid of blank columns
|
|
|
139 for (my $j=0;$j<scalar @cols;$j++){
|
|
|
140 $cols[$j]=$cols[$j].'-'.$labels[0]; #mark each column name with file it comes from
|
|
|
141 }
|
|
|
142 push (@header,@cols);
|
|
|
143
|
|
|
144 while (my $line=<F1>){
|
|
|
145 chomp $line;
|
|
|
146 my @info=split(",",$line);
|
|
|
147 #Only keep the first 7 columns (Ones about blanks aren't needed for comparisons)
|
|
|
148 @info=@info[0,1,2,3,4,5,6];
|
|
|
149 #Sometimes genes that don't have a gene name can't be blank, so fill with NA
|
|
|
150 if (!$info[5]){
|
|
|
151 $info[5]="NA";
|
|
|
152 }
|
|
|
153 #If there are no insertions in the column "total", then make it =0 rather than blank
|
|
|
154 if (!$info[6]){
|
|
|
155 $info[6]=0;
|
|
|
156 }
|
|
|
157 $one{$info[0]}=\@info;
|
|
|
158 }
|
|
|
159 close F1;
|
|
|
160
|
|
|
161 #FILE2 OPENING ---> %two WHERE KEY:VALUE IS GENE_ID:(GENE_ID,INSERTIONS,MEAN,ETC.)
|
|
|
162
|
|
|
163 my %two;
|
|
|
164 open (F2,'<',$files[1]);
|
|
|
165
|
|
|
166 #STORE COLUMN NAMES (FIRST LINE OF FILE2) FOR HEADER AND APPEND LABELS
|
|
|
167 $head=<F2>; #the header in the file
|
|
|
168 @cols=split(',',$head);
|
|
|
169 @cols=@cols[0,1,2,3,4,5,6]; #get rid of blank columns
|
|
|
170 for (my $j=0;$j<scalar @cols;$j++){
|
|
|
171 $cols[$j]=$cols[$j].'-'.$labels[1]; #mark each column name with file it comes from
|
|
|
172 }
|
|
|
173 push (@header,@cols);
|
|
|
174
|
|
|
175 while (my $line=<F2>){
|
|
|
176 chomp $line;
|
|
|
177 my @info=split(",",$line);
|
|
|
178 @info=@info[0,1,2,3,4,5,6];
|
|
|
179 if (!$info[5]){
|
|
|
180 $info[5]="NA";
|
|
|
181 }
|
|
|
182 if (!$info[6]){
|
|
|
183 $info[6]=0;
|
|
|
184 }
|
|
|
185 $two{$info[0]}=\@info;
|
|
|
186 }
|
|
|
187 close F2;
|
|
|
188
|
|
|
189
|
|
|
190 #READ CONVERSION FILE INTO ARRAY.
|
|
|
191 #Conversion file must have strain 1 for file 1 in column 1 (index 0) and
|
|
|
192 #strain 2 for file 2 in column 2 (index 1)
|
|
|
193 #conversion file must be tab delimited with no NA fields
|
|
|
194 #If homologs (exist then take info from hashes (%one and %two) by referring to gene_id in KEY
|
|
|
195
|
|
|
196 my @all; #store all homologs in this hash
|
|
|
197 open (CONV,'<',$cfile);
|
|
|
198 while (my $line=<CONV>){
|
|
|
199 chomp $line;
|
|
|
200 my @genes=split("\t",$line); #Array @genes will contain two genes (SP_0000,SPT_0000)
|
|
|
201 if (scalar @genes==2 and $genes[0] ne "" and $genes[1] ne ""){
|
|
|
202 my @info;
|
|
|
203 my @oneArray=@{$one{$genes[0]}};
|
|
|
204 my @twoArray=@{$two{$genes[1]}};
|
|
|
205 push (@info,@oneArray,@twoArray);
|
|
|
206 my $diff=sprintf("$round",($info[1]-$info[8]));
|
|
|
207 my $total1=$info[6];
|
|
|
208 my $total2=$info[13];
|
|
|
209 my $sd1=$info[3];
|
|
|
210 my $se1=$info[4];
|
|
|
211 my $sd2=$info[10];
|
|
|
212 my $se2=$info[11];
|
|
|
213 my $df=$total1+$total2-2;
|
|
|
214 my $tdist;
|
|
|
215 my $pval;
|
|
|
216 #TDIST, PVAL calculations with fail if standard dev, error, or counts are not real numbers
|
|
|
217 #or if 0 ends up in denominator
|
|
|
218 if ($se1 eq "X" or $se2 eq "X" or $sd1 eq "X" or $sd2 eq "X" or $total1==0 or $total2==0 or $sd1==0 or $sd2==0){
|
|
|
219 ($tdist,$pval)=("NA","NA");
|
|
|
220 }
|
|
|
221 else{
|
|
|
222 $tdist=sqrt((($diff)/(sqrt((($sd1**2)/$total1)+(($sd2**2)/$total2))))**2);
|
|
|
223 $pval=Statistics::Distributions::tprob($df,$tdist);
|
|
|
224 }
|
|
|
225 push (@info,$diff,$df,$tdist,$pval);
|
|
|
226 push (@all,\@info);
|
|
|
227 }
|
|
|
228 }
|
|
|
229 close CONV;
|
|
|
230
|
|
|
231 #SORT THE HOMOLOGS BY THE SORTKEY OR BY DEFAULT DIFFERENCE IN MEAN FITNESSES
|
|
|
232 if (!$sortkey){
|
|
|
233 $sortkey=14; #for mean difference
|
|
|
234 }
|
|
|
235 my @sorted = sort { $b->[$sortkey] <=> $a->[$sortkey] } @all;
|
|
|
236
|
|
|
237 #FINISH THE HEADER BY ADDING COLUMN NAMES FOR MEAN-DIFF, DOF, TDIST, AND PVALUE
|
|
|
238 my $field="MeanDiff(".$labels[0].'.'.$labels[1].")";
|
|
|
239 push (@header,$field,"DOF","TDIST","PVALUE");
|
|
|
240
|
|
|
241 #PRINT MATCHED HOMOLOG INFORMATION INTO A SINGLE OUTPUT FILE
|
|
|
242 open OUT, '>',"$out";
|
|
|
243 print OUT (join(',',@header),"\n");
|
|
|
244 foreach (@sorted){
|
|
|
245 my @woo=@{$_};
|
|
|
246 print OUT join(',',@woo),"\n";
|
|
|
247 }
|
|
|
248
|
|
|
249 close OUT;
|
|
|
250
|
|
|
251
|
