| 1 |
#!/usr/bin/perl |
| 2 |
use strict; |
| 3 |
use Getopt::Std; |
| 4 |
use FindBin;use lib $FindBin::Bin; |
| 5 |
|
| 6 |
my $usage = q/Usage: |
| 7 |
tbl2gff [-F] [-g <chr_table>] [-t <track>] -p <prot2nuc.tab> <ncbi_tbl_file> |
| 8 |
|
| 9 |
Convert the given NCBI tbl file to GFF3 |
| 10 |
|
| 11 |
Options: |
| 12 |
-p protein_id to transcript_id mapping file (required) |
| 13 |
this file has 2 columns: |
| 14 |
GenPept accession and (nucleotide) accession |
| 15 |
it is required in order to assign CDS to their |
| 16 |
corresponding mRNA exons |
| 17 |
-t use given track name instead of 'gb_tbl' |
| 18 |
-g replace the genomic sequence names according |
| 19 |
to the translation table <chr_table> |
| 20 |
(only the first 2 columns in that file are used) |
| 21 |
-M only outputs mRNA entries (without gene lines) |
| 22 |
-F preserve all notes and comments |
| 23 |
/; |
| 24 |
umask 0002; |
| 25 |
my %kattrs; |
| 26 |
#key attributes to always keep |
| 27 |
@kattrs{'gene', 'product', 'transcript_id'}=(); |
| 28 |
getopts('FMt:g:p:') || die($usage."\n"); |
| 29 |
my $gxfile=$Getopt::Std::opt_g; |
| 30 |
my $p2nfile=$Getopt::Std::opt_p; |
| 31 |
my $mrnaonly=$Getopt::Std::opt_M; |
| 32 |
die("Error: -p is required -- to map protein_id to transcript_id\n") unless $p2nfile && -f $p2nfile; |
| 33 |
my %pmap; # protein_id => transcript_id |
| 34 |
open(PROTMAP, $p2nfile) || die("Error opening $p2nfile!\n"); |
| 35 |
while (<PROTMAP>) { |
| 36 |
chomp; |
| 37 |
my ($p, $n)=split; |
| 38 |
next unless $n; |
| 39 |
$p=~s/\.\d+$//; |
| 40 |
$n=~s/\.\d+$//; # discard versioning |
| 41 |
$pmap{$p}=$n; |
| 42 |
} |
| 43 |
close(PROTMAP); |
| 44 |
|
| 45 |
my $track=$Getopt::Std::opt_t || 'gb_tbl'; |
| 46 |
my $allattrs=$Getopt::Std::opt_F; |
| 47 |
my $last_line; |
| 48 |
my %gx; # translation table for genomic sequence names ('>Feature' line) |
| 49 |
if ($gxfile) { |
| 50 |
open(GX, $gxfile) || die("Error opening $gxfile\n"); |
| 51 |
while(<GX>) { |
| 52 |
chomp; |
| 53 |
my @t=split; |
| 54 |
if ($t[0] && $t[1]) { |
| 55 |
$gx{$t[0]}=$t[1]; |
| 56 |
$gx{$t[1]}=$t[0]; |
| 57 |
} |
| 58 |
} |
| 59 |
close(GX); |
| 60 |
} |
| 61 |
|
| 62 |
my $tblfile=$ARGV[0]; |
| 63 |
die($usage."\n") if $tblfile eq '-h' || ! -f $tblfile; |
| 64 |
|
| 65 |
my %gids; #ensure unique gene IDs |
| 66 |
#allow for duplication $gids{$gene}=[ [gstart1, gend1], [gstart2, gend2] ] |
| 67 |
my %tids; # and transcript IDs |
| 68 |
|
| 69 |
# -- per gbase: |
| 70 |
|
| 71 |
my %genes; # genes{$gene}=[$gbase, $strand, $gstart, $gend, [@gxrefs], \%gattrs, \%rnas, isPseudo] |
| 72 |
# 0 1 2 3 4 5 6 7 |
| 73 |
#my %rnas; |
| 74 |
# %rnas hash holds all mRNA data for a single gene (i.e. all isoforms) |
| 75 |
# 0 1 2 3 4 5 6 |
| 76 |
# $rnas{transcript_id}=[ $featname, [@exon], [@cds], \%tattrs, [@txrefs], $partial3, $partial5 ] |
| 77 |
#current values: |
| 78 |
my ($gbase, $curfeat, @exon, %cattrs, $gpseudo, $gstart, $gend, $partial5, $partial3); |
| 79 |
my $strand='+'; |
| 80 |
|
| 81 |
# product and transcript_id are found in %tattrs |
| 82 |
while (<>) { |
| 83 |
$last_line=$_; |
| 84 |
if (m/^>Feature\s+(\S+)/) { # misnomer - in fact this is the genomic sequence name |
| 85 |
my $newbase=$1; |
| 86 |
flushGSeq(); #write previous gene feature |
| 87 |
$gbase=''; |
| 88 |
if ($gxfile) { |
| 89 |
$gbase=$gx{$newbase}; |
| 90 |
unless ($gbase) { |
| 91 |
if ($newbase=~m/^[a-z]+\|([\w\.]+)/) { |
| 92 |
$newbase=$1; |
| 93 |
$gbase=$gx{$newbase}; |
| 94 |
unless ($gbase) { |
| 95 |
$newbase=~s/\.\d+$//; |
| 96 |
$gbase=$gx{$newbase}; |
| 97 |
} |
| 98 |
} |
| 99 |
} |
| 100 |
} #convert contig name |
| 101 |
$gbase=$newbase unless $gbase; |
| 102 |
next; |
| 103 |
} # base genomic sequence (contig/chromosome) established |
| 104 |
chomp; |
| 105 |
tr/\n\r//d; |
| 106 |
my @t=split(/\t/); |
| 107 |
my ($start, $end)=@t[0,1]; |
| 108 |
if ($start && $end) { # segment coordinates |
| 109 |
$partial5=1 if ($start=~s/^<//); |
| 110 |
$partial3=1 if ($end=~s/^>//); |
| 111 |
my $newstrand='+'; |
| 112 |
if ($end<$start) { |
| 113 |
$newstrand='-'; |
| 114 |
($start, $end)=($end, $start); |
| 115 |
} |
| 116 |
if ($t[2]) { # feature start (gene/mRNA/CDS) |
| 117 |
flushFeature() unless $t[2] eq 'exon'; #cleanup current gene/transcript data |
| 118 |
$strand=$newstrand; |
| 119 |
if ($t[2] eq 'gene') { |
| 120 |
#new gene feature starts |
| 121 |
$curfeat='gene'; |
| 122 |
$gstart=$start; |
| 123 |
$gend=$end; |
| 124 |
} # gene start |
| 125 |
else { #non-gene features starting |
| 126 |
if ($t[2] eq 'exon') { # special case, exons given one by one |
| 127 |
if ($curfeat ne 'RNA_exon') { #first exon |
| 128 |
flushFeature(); #just in case there was another mRNA (exons) given before |
| 129 |
} |
| 130 |
$curfeat='RNA_exon'; |
| 131 |
push(@exon, [$start, $end]); |
| 132 |
} |
| 133 |
elsif ($t[2] eq 'CDS' || $t[2]=~m/RNA/) { |
| 134 |
$curfeat=$t[2]; |
| 135 |
push(@exon, [$start, $end]); |
| 136 |
} |
| 137 |
else { $curfeat=''; next; } |
| 138 |
# -- skipping any other features here |
| 139 |
} #non-gene features starting |
| 140 |
} # feature start |
| 141 |
else { # feature continuation |
| 142 |
$strand=$newstrand; |
| 143 |
if ($curfeat eq 'CDS' || $curfeat=~/RNA/) { |
| 144 |
push(@exon, [$start, $end]); |
| 145 |
} |
| 146 |
else { |
| 147 |
die("Error: unexpected feature '$curfeat' continuation\n$last_line"); |
| 148 |
} |
| 149 |
} |
| 150 |
} # had segment coordinates |
| 151 |
else { # no segment coords => attributes for current feature |
| 152 |
my ($attr, $value)=@t[3,4]; |
| 153 |
$value=~tr/;/,/; |
| 154 |
if ($attr eq 'db_xref' || $attr eq 'gene_syn') { |
| 155 |
push(@{$cattrs{$attr}}, $value); |
| 156 |
} |
| 157 |
else { |
| 158 |
$cattrs{$attr}=$value if $attr ne 'number'; |
| 159 |
} |
| 160 |
} # attributes block |
| 161 |
} #while input lines |
| 162 |
|
| 163 |
flushGSeq(); |
| 164 |
|
| 165 |
#================ SUBROUTINES ============ |
| 166 |
|
| 167 |
sub writeGene { |
| 168 |
my ($gene, $gdata)=@_; |
| 169 |
my $geneid=$gene; |
| 170 |
my ($gene_name)=($geneid=~m/^([^~]+)/); |
| 171 |
$geneid=~tr/~/_/; |
| 172 |
($gbase, $strand, $gstart, $gend, $gpseudo)= |
| 173 |
($gdata->[0], $gdata->[1], $gdata->[2], $gdata->[3], $gdata->[7]); |
| 174 |
my $gfeature=($gpseudo)?'pseudogene' : 'gene'; |
| 175 |
#makeUniqID(\$gene, \%gids); # $gene will be updated if necessary |
| 176 |
#my $geneid=$gene; |
| 177 |
#$geneid=~tr/~/_/; |
| 178 |
my $attrs="ID=$geneid;Name=$gene_name"; |
| 179 |
my @gxrefs=@{$gdata->[4]}; |
| 180 |
$attrs.=';xrefs='.join(',', @gxrefs) if (@gxrefs>0); |
| 181 |
my $gattrs=$gdata->[5]; |
| 182 |
unless ($mrnaonly) { |
| 183 |
foreach my $k (keys(%$gattrs)) { |
| 184 |
$attrs.=";$k=".$gattrs->{$k}; |
| 185 |
} |
| 186 |
print join("\t", $gbase, $track, $gfeature, $gstart, $gend, '.', $strand, '.', $attrs)."\n"; |
| 187 |
} |
| 188 |
|
| 189 |
my $rnas=$gdata->[6]; |
| 190 |
foreach my $tid (keys(%$rnas)) { |
| 191 |
#die("Error: no transcript_id found at write_RNA for gene $gene ($geneid)\n$last_line\n") unless $tid; |
| 192 |
my $tdata=$rnas->{$tid}; |
| 193 |
my $rnafeat=$tdata->[0]; |
| 194 |
if ($rnafeat eq 'RNA_exon') { |
| 195 |
$rnafeat = $gpseudo ? 'pseudo_RNA' : 'misc_RNA'; |
| 196 |
} |
| 197 |
if ($mrnaonly) { |
| 198 |
$rnafeat=~s/misc_RNA/mRNA/; |
| 199 |
} |
| 200 |
|
| 201 |
next if $mrnaonly && $rnafeat ne 'mRNA'; |
| 202 |
my $texons=$tdata->[1]; |
| 203 |
my $tcds=$tdata->[2]; |
| 204 |
my $tahref=$tdata->[3]; #attributes |
| 205 |
my $txar=$tdata->[4]; # db_xrefs |
| 206 |
my ($part3, $part5)=($tdata->[5], $tdata->[6]); |
| 207 |
makeUniqID(\$tid, \%tids); |
| 208 |
@exon=sort { $main::a->[0] <=> $main::b->[0] } @$texons; |
| 209 |
my ($tstart, $tend)=($exon[0]->[0], $exon[-1]->[1]); |
| 210 |
my $attrs="ID=$tid"; |
| 211 |
$attrs.=";Parent=$geneid" unless $mrnaonly; |
| 212 |
$attrs.="Name=$gene_name;gene_name=$gene_name"; |
| 213 |
$attrs.=';partial=1' if ($part5 || $part3); |
| 214 |
my $product=$tahref->{'product'}; |
| 215 |
if ($product) { |
| 216 |
$product=~tr/"/'/; #" |
| 217 |
$attrs.=';product="'.$product.'"'; |
| 218 |
} |
| 219 |
if ($allattrs) { |
| 220 |
$attrs.=';xrefs='.join(',', @$txar) if (@$txar > 0); |
| 221 |
foreach my $k (keys(%$tahref)) { |
| 222 |
next if exists($kattrs{$k}); |
| 223 |
$attrs.=";$k=".$tahref->{$k}; |
| 224 |
} |
| 225 |
} |
| 226 |
print join("\t", $gbase, $track, $rnafeat, $tstart, $tend, '.', $strand, '.', $attrs)."\n"; |
| 227 |
foreach my $ex (@exon) { |
| 228 |
print join("\t", $gbase, $track, 'exon', $$ex[0], $$ex[1], '.', $strand, '.', "Parent=$tid")."\n"; |
| 229 |
} |
| 230 |
@exon=(); |
| 231 |
#write_CDS(); |
| 232 |
my @cds=@$tcds; |
| 233 |
if (@cds>0) { #have CDS, write it |
| 234 |
@cds=sort { $main::a->[0] <=> $main::b->[0] } @cds; |
| 235 |
my $codonstart=$tahref->{'codon_start'}; |
| 236 |
if ($codonstart>1) { |
| 237 |
$codonstart--; |
| 238 |
if ($strand eq '-') { |
| 239 |
$cds[-1]->[1]-=$codonstart; |
| 240 |
} |
| 241 |
else { |
| 242 |
$cds[0]->[0]+=$codonstart; |
| 243 |
} |
| 244 |
} |
| 245 |
else { $codonstart='0'; } |
| 246 |
|
| 247 |
foreach my $cd (@cds) { #TODO: calculate phase ? |
| 248 |
print join("\t", $gbase, $track, 'CDS', $$cd[0], $$cd[1], '.', $strand, '.', "Parent=$tid")."\n"; |
| 249 |
} |
| 250 |
} |
| 251 |
} #for each transcript |
| 252 |
} |
| 253 |
|
| 254 |
sub makeUniqID { |
| 255 |
my ($idref, $href)=@_; |
| 256 |
my $id=$$idref; |
| 257 |
$id=~s/~\d+$//; |
| 258 |
my $c = ++$$href{$id}; |
| 259 |
if ($c>1) { |
| 260 |
$c--; |
| 261 |
$$idref=$id.'.m'.$c; |
| 262 |
} |
| 263 |
return $$idref; |
| 264 |
} |
| 265 |
|
| 266 |
sub flushFeature { #called after an attribute block has been parsed into %cattrs |
| 267 |
return unless $curfeat && keys(%cattrs)>0; |
| 268 |
my $gene=$cattrs{'gene'}; |
| 269 |
die("Error: no gene attribute found for current feature $curfeat\n") |
| 270 |
unless $gene; |
| 271 |
if ($curfeat eq 'gene') { # gene attributes processing |
| 272 |
#make sure it's unique |
| 273 |
push(@{$gids{$gene}},[$gbase, $gstart, $gend]); |
| 274 |
my $gno=scalar(@{$gids{$gene}})-1; |
| 275 |
$gene.='~'.$gno if $gno>0; |
| 276 |
if (exists($genes{$gene})) { #should never happen, we just made sure it's unique! |
| 277 |
my $prev_gstart=$genes{$gene}->[2]; |
| 278 |
my $prev_gend=$genes{$gene}->[3]; |
| 279 |
# if (abs($prev_gstart-$gstart)>30000) { |
| 280 |
die("Error -- gene $gene is duplicated ($prev_gstart vs $gstart) !?\n"); |
| 281 |
# } |
| 282 |
# else { #update min..max coordinates |
| 283 |
# $genes{$gene}->[2]=$gstart if $gstart<$prev_gstart; |
| 284 |
# $genes{$gene}->[3]=$gend if $gend>$prev_gend; |
| 285 |
# ${$genes{$gene}->[5]}{'fragmented'}=1; |
| 286 |
# } |
| 287 |
} |
| 288 |
else |
| 289 |
{ # create a new gene entry here |
| 290 |
my $pseudo= delete($cattrs{'pseudo'}) ? 1 : 0; |
| 291 |
# 0 1 2 3 4 5 6 7 |
| 292 |
$genes{$gene}=[$gbase, $strand, $gstart, $gend, [], {}, {}, $pseudo]; |
| 293 |
} |
| 294 |
delete $cattrs{'gene'}; |
| 295 |
if ($allattrs) { |
| 296 |
if (exists($cattrs{'db_xref'})) { |
| 297 |
$genes{$gene}->[4]=[@{$cattrs{'db_xref'}}]; |
| 298 |
delete $cattrs{'db_xref'}; |
| 299 |
} |
| 300 |
#another special case: gene_syn |
| 301 |
if (exists($cattrs{'gene_syn'})) { |
| 302 |
#convert [list] to a string with comma delimited names |
| 303 |
$cattrs{'gene_syn'}=join(',',@{$cattrs{'gene_syn'}}); |
| 304 |
} |
| 305 |
my $garef=$genes{$gene}->[5]; |
| 306 |
my @k=keys(%cattrs); |
| 307 |
@{$garef}{@k}=(values(%cattrs)); |
| 308 |
} #all attributes are kept |
| 309 |
@exon=(); |
| 310 |
%cattrs=(); |
| 311 |
($partial3, $partial5)=(); |
| 312 |
return; #gene data added |
| 313 |
} # stored gene attributes |
| 314 |
# else |
| 315 |
# - storing RNA attributes |
| 316 |
# now make sure we place this RNA within the right gene - in case of gene duplications |
| 317 |
# locate the matching $gene |
| 318 |
@exon= sort { $main::a->[0] <=> $main::b->[0] } @exon; |
| 319 |
my ($xstart, $xend)=($exon[0]->[0], $exon[-1]->[1]); |
| 320 |
my $gxdata=$gids{$gene}; |
| 321 |
die("Error: gene $gene not found in \%gids !\n") unless $gxdata; #should never happen |
| 322 |
if (@$gxdata>1) { |
| 323 |
my $i=0; |
| 324 |
my $found; |
| 325 |
foreach my $gd (@$gxdata) { |
| 326 |
if ($gd->[0] eq $gbase && $gd->[1]<=$xstart && $gd->[2]>=$xend) { |
| 327 |
$found=1; |
| 328 |
last; |
| 329 |
} |
| 330 |
$i++; |
| 331 |
} |
| 332 |
die("Error: cannot find gene enclosing RNA data ". |
| 333 |
$cattrs{'transcript_id'}." ".$cattrs{'protein_id'}." (on $gbase, $xstart..$xend)\n") |
| 334 |
unless $found; |
| 335 |
$gene.='~'.$i if $i>0; |
| 336 |
} |
| 337 |
|
| 338 |
my $geneid=$gene; |
| 339 |
my $gdata=$genes{$gene}; |
| 340 |
die("Error: \$genes data for gene '$gene' cannot be found!\n") unless $gdata; |
| 341 |
my $rnas=$gdata->[6]; |
| 342 |
$geneid=~tr/~/_/; |
| 343 |
my ($tid, $pid); |
| 344 |
if ($curfeat eq 'CDS') { # just finished a CDS parsing |
| 345 |
my $pid=$cattrs{'protein_id'}; |
| 346 |
die("Error: no protein_id found for CDS feature (gene $geneid)!\n") unless $pid; |
| 347 |
#$pid=~s/^[a-z]+\|//i; # ref|NM_002314.1| |
| 348 |
#$pid=~s/\|(\w*)$//; # some entries have a suffix - e.g. ref|NM_006824.1|P40 |
| 349 |
my @t=split(/\|/,$pid); |
| 350 |
$pid=$t[1] if (@t>1); |
| 351 |
$pid=~s/\.\d+$//; #remove version |
| 352 |
$tid=$pmap{$pid}; |
| 353 |
die("Error: cannot get transcript for protein id $pid (gene $geneid)!\n") unless $tid; |
| 354 |
unless (exists($rnas->{$tid})) { #this should never happen - CDS without mRNA definition? |
| 355 |
# 0 1 2 3 4 5 6 |
| 356 |
#$rnas->{$tid}=['mRNA', [@exons], [@cds], {}, [], 0, 0]; |
| 357 |
$tid=~s/\.\d+$//; |
| 358 |
die("Error: cannot find mRNA entry for protein $pid (reported as coming from transcript $tid)\n") |
| 359 |
unless exists($rnas->{$tid}); |
| 360 |
} |
| 361 |
# #just add the CDS |
| 362 |
push(@{$rnas->{$tid}->[2]}, @exon); |
| 363 |
# } |
| 364 |
} #CDS |
| 365 |
else { # exons (RNA feature) |
| 366 |
$tid=$cattrs{'transcript_id'}; |
| 367 |
if ($tid) { |
| 368 |
my @t=split(/\|/,$tid); |
| 369 |
$tid=$t[1] if (@t>1); |
| 370 |
#$tid=~s/^[a-z]+\|//i; |
| 371 |
#$tid=~s/\|\w*$//; # some entries have a suffix - e.g. ref|NM_006824.1|P40 |
| 372 |
$tid=~s/\.\d+$//; #remove version! |
| 373 |
} |
| 374 |
else { |
| 375 |
$tid=$geneid.'_t'; #some special cases with 'exon' features for pseudo genes |
| 376 |
} |
| 377 |
unless (exists($rnas->{$tid})) { |
| 378 |
# 0 1 2 3 4 5 6 |
| 379 |
$rnas->{$tid}=[$curfeat, [@exon], [], {}, [], 0, 0]; |
| 380 |
} |
| 381 |
else { #just add the exons |
| 382 |
push(@{$rnas->{$tid}->[1]}, @exon); |
| 383 |
} |
| 384 |
} #exons |
| 385 |
my $tdata=$rnas->{$tid}; |
| 386 |
$tdata->[5]=1 if $partial5; |
| 387 |
$tdata->[6]=1 if $partial3; |
| 388 |
# add attributes |
| 389 |
delete @cattrs{'gene', 'transcript_id', 'protein_id'}; |
| 390 |
my $product = delete $cattrs{'product'}; |
| 391 |
my $ahref=$tdata->[3]; |
| 392 |
$ahref->{'product'}=$product if $product; |
| 393 |
if ($allattrs) { |
| 394 |
my %xrefs; |
| 395 |
my @cx; |
| 396 |
@cx=@{$cattrs{'db_xref'}} if exists $cattrs{'db_xref'}; |
| 397 |
push(@cx, @{$tdata->[4]}); |
| 398 |
@xrefs{@cx}=(); |
| 399 |
@cx=keys(%xrefs); |
| 400 |
$tdata->[4]=[@cx]; |
| 401 |
delete $cattrs{'db_xref'}; |
| 402 |
#add new attributes |
| 403 |
foreach my $k (keys %cattrs) { |
| 404 |
$ahref->{$k}=$cattrs{$k}; |
| 405 |
} |
| 406 |
} |
| 407 |
#--> clean up |
| 408 |
@exon=(); |
| 409 |
%cattrs=(); |
| 410 |
($partial3, $partial5)=(); |
| 411 |
} |
| 412 |
|
| 413 |
sub flushGSeq { |
| 414 |
return unless keys(%genes); |
| 415 |
flushFeature(); # write last RNA feature read |
| 416 |
my $gcount=keys(%genes); |
| 417 |
while (my ($g,$gd)=each(%genes)) { |
| 418 |
writeGene($g,$gd); |
| 419 |
} |
| 420 |
($gbase, $curfeat, @exon, %cattrs, $gpseudo, $gstart, $gend, |
| 421 |
$partial5, $partial3)=(); |
| 422 |
$strand='+'; |
| 423 |
@exon=(); |
| 424 |
%cattrs=(); |
| 425 |
%genes=(); |
| 426 |
#-- debug only: |
| 427 |
#exit; |
| 428 |
} |
