| 1 |
package Genome::Model::Tools::Music::Proximity; |
| 2 |
|
| 3 |
use warnings; |
| 4 |
use strict; |
| 5 |
use IO::File; |
| 6 |
|
| 7 |
our $VERSION = $Genome::Model::Tools::Music::VERSION; |
| 8 |
|
| 9 |
class Genome::Model::Tools::Music::Proximity { |
| 10 |
is => 'Command::V2', |
| 11 |
has_input => [ |
| 12 |
maf_file => { is => 'Text', doc => "List of mutations using TCGA MAF specifications v2.3" }, |
| 13 |
output_dir => { is => 'Text', doc => "Directory where output files will be written" }, |
| 14 |
max_proximity => { is => 'Text', doc => "Maximum allowed AA distance between 2 mutations", is_optional => 1, default => 7 }, |
| 15 |
skip_non_coding => { is => 'Boolean', doc => "Skip non-coding mutations from the provided MAF file", is_optional => 1, default => 1 }, |
| 16 |
skip_silent => { is => 'Boolean', doc => "Skip silent mutations from the provided MAF file", is_optional => 1, default => 1 }, |
| 17 |
], |
| 18 |
has_output => [ |
| 19 |
output_file => {is => 'Text', doc => "TODO"}, |
| 20 |
], |
| 21 |
doc => "Perform a proximity analysis on a list of mutations." |
| 22 |
}; |
| 23 |
|
| 24 |
sub help_detail { |
| 25 |
return <<HELP |
| 26 |
This module first calculates the amino acid position of each mutation in the MAF file within its |
| 27 |
respective transcript. Then, for each mutation, two values are calculated: 1) the number of other |
| 28 |
mutations on the same transcript within the proximity limit set by the max-proximity input |
| 29 |
parameter, and 2) the distance to the closest other mutation in this nearby set. Only mutations |
| 30 |
which have another mutation within close proximity are reported in the output-file. |
| 31 |
|
| 32 |
In addition to the standard version 2.3 MAF headers, there needs to be 3 columns appended. These |
| 33 |
column headers in the MAF must have these names in the header in order for the tool to find them: |
| 34 |
transcript_name - the transcript name, such as NM_000028 |
| 35 |
amino_acid_change - the amino acid change, such as p.R290H |
| 36 |
c_position - the nucleotide position changed, such as c.869 |
| 37 |
|
| 38 |
The output is generated with the folowing column headers: |
| 39 |
Mutations_Within_Proximity, Nearest_Mutation, Gene, Transcript, Affected_Amino_Acid(s), Chr, |
| 40 |
Start, Stop, Ref_Allele, Var_Allele, Sample |
| 41 |
HELP |
| 42 |
} |
| 43 |
|
| 44 |
sub help_synopsis { |
| 45 |
return <<EOS |
| 46 |
... music proximity \\ |
| 47 |
--maf-file input_dir/myMAF.tsv \\ |
| 48 |
--output-dir output_dir/ \\ |
| 49 |
--max-proximity 15 |
| 50 |
EOS |
| 51 |
} |
| 52 |
|
| 53 |
sub _doc_authors { |
| 54 |
return <<EOS |
| 55 |
Nathan D. Dees, Ph.D. |
| 56 |
Dan Koboldt, M.S. |
| 57 |
Cyriac Kandoth, Ph.D. |
| 58 |
EOS |
| 59 |
} |
| 60 |
|
| 61 |
sub execute { |
| 62 |
my $self = shift; |
| 63 |
my $maf_file = $self->maf_file; |
| 64 |
my $output_dir = $self->output_dir; |
| 65 |
my $max_proximity = $self->max_proximity; |
| 66 |
my $skip_non_coding = $self->skip_non_coding; |
| 67 |
my $skip_silent = $self->skip_silent; |
| 68 |
$output_dir =~ s/(\/)+$//; # Remove trailing forward slashes if any |
| 69 |
|
| 70 |
# Check on all the input data before starting work |
| 71 |
print STDERR "MAF file not found or is empty: $maf_file\n" unless( -s $maf_file ); |
| 72 |
print STDERR "Output directory not found: $output_dir\n" unless( -e $output_dir ); |
| 73 |
return undef unless( -s $maf_file && -e $output_dir ); |
| 74 |
|
| 75 |
# Output of this script will be written to this location in the output directory |
| 76 |
my $out_file = "$output_dir/proximity_report"; |
| 77 |
$self->output_file($out_file); |
| 78 |
|
| 79 |
# Parse the header row in the MAF file |
| 80 |
my $maf_fh = IO::File->new( $maf_file ) or die "Couldn't open $maf_file. $!"; |
| 81 |
my $maf_header = $maf_fh->getline; |
| 82 |
$maf_header = $maf_fh->getline while( $maf_header =~ /^#/ ); # Skip commented lines |
| 83 |
chomp( $maf_header ); |
| 84 |
unless( $maf_header =~ /^Hugo_Symbol/ ) |
| 85 |
{ |
| 86 |
print STDERR "Could not find column headers in $maf_file\n"; |
| 87 |
return undef; |
| 88 |
} |
| 89 |
|
| 90 |
# Check whether the required additional MAF columns were included in the MAF |
| 91 |
unless( $maf_header =~ m/c_position/ and $maf_header =~ m/amino_acid_change/ and $maf_header =~ m/transcript_name/ ) |
| 92 |
{ |
| 93 |
print STDERR "Could not find required additional columns in $maf_file\n"; |
| 94 |
return undef; |
| 95 |
} |
| 96 |
|
| 97 |
# Find the indexes of all the MAF columns |
| 98 |
my $idx = 0; |
| 99 |
my %col_idx = map {($_, $idx++)} split( /\t/, $maf_header ); |
| 100 |
|
| 101 |
# A hash to store statuses, and a hash to store variants and AA positions |
| 102 |
my %status; |
| 103 |
my $status = \%status; |
| 104 |
my %aa_mutations; |
| 105 |
|
| 106 |
# Load relevant data from MAF into hash |
| 107 |
while( my $line = $maf_fh->getline ) |
| 108 |
{ |
| 109 |
chomp $line; |
| 110 |
my @cols = split( /\t/, $line ); |
| 111 |
|
| 112 |
# Fetch data from the generic MAF columns |
| 113 |
my ( $gene, $chr, $start, $stop, $mutation_class, $mutation_type, $ref_allele, $var_allele, $var2, $sample ) = |
| 114 |
( $cols[0], $cols[4], $cols[5], $cols[6], $cols[8], $cols[9], $cols[10], $cols[11], $cols[12], $cols[15] ); |
| 115 |
$var_allele = $var2 if ( $var_allele eq $ref_allele ); # Different centers interpret the 2 variant columns differently |
| 116 |
|
| 117 |
# Fetch data from the required additional MAF columns |
| 118 |
my ( $c_position, $aa_change, $transcript ) = ( $cols[$col_idx{c_position}], $cols[$col_idx{amino_acid_change}], $cols[$col_idx{transcript_name}] ); |
| 119 |
|
| 120 |
# Create a key to uniquely identify each variant |
| 121 |
my $variant_key = join( "\t", $gene, $chr, $start, $stop, $ref_allele, $var_allele, $sample ); |
| 122 |
|
| 123 |
|
| 124 |
#check that the mutation class is acceptable |
| 125 |
if( $mutation_class !~ m/^(Missense_Mutation|Nonsense_Mutation|Nonstop_Mutation|Splice_Site|Translation_Start_Site|Frame_Shift_Del|Frame_Shift_Ins|In_Frame_Del|In_Frame_Ins|Silent|Intron|RNA|3'Flank|3'UTR|5'Flank|5'UTR|IGR|Targeted_Region|De_novo_Start_InFrame|De_novo_Start_OutOfFrame)$/ ) |
| 126 |
{ |
| 127 |
print STDERR "Unrecognized Variant_Classification \"$mutation_class\" in MAF file for gene $gene\n"; |
| 128 |
print STDERR "Please use TCGA MAF Specification v2.3.\n"; |
| 129 |
return undef; |
| 130 |
} |
| 131 |
|
| 132 |
# If user wants, skip Silent mutations, or those in Introns, RNA, UTRs, Flanks, IGRs, or the ubiquitous Targeted_Region |
| 133 |
if(( $skip_non_coding && $mutation_class =~ m/^(Intron|RNA|3'Flank|3'UTR|5'Flank|5'UTR|IGR|Targeted_Region)$/ ) || |
| 134 |
( $skip_silent && $mutation_class =~ m/^Silent$/ )) |
| 135 |
{ |
| 136 |
print "Skipping $mutation_class mutation in gene $gene.\n"; |
| 137 |
$status{synonymous_mutations_skipped}++; |
| 138 |
next; |
| 139 |
} |
| 140 |
|
| 141 |
# Determine amino acid position and load into hash |
| 142 |
my @mutated_aa_positions = (); |
| 143 |
@mutated_aa_positions = $self->get_amino_acid_pos( $variant_key, $mutation_class, $c_position, $aa_change, $status ); |
| 144 |
|
| 145 |
# Record data in hash if mutated aa position found |
| 146 |
if( scalar( @mutated_aa_positions ) > 0 ) |
| 147 |
{ |
| 148 |
push( @{$aa_mutations{$transcript}{$variant_key}{mut_AAs}}, @mutated_aa_positions ); |
| 149 |
} |
| 150 |
} |
| 151 |
$maf_fh->close; |
| 152 |
|
| 153 |
# Evaluate proximity of mutated amino acids for each transcript |
| 154 |
for my $transcript ( keys %aa_mutations ) |
| 155 |
{ |
| 156 |
# For each variant hitting that transcript |
| 157 |
for my $variant ( keys %{$aa_mutations{$transcript}} ) |
| 158 |
{ |
| 159 |
# Initialize the search |
| 160 |
my @affected_amino_acids = @{$aa_mutations{$transcript}{$variant}{mut_AAs}}; |
| 161 |
my $mutations_within_proximity = 0; # Variable for summing # of mutations within proximity |
| 162 |
my $min_proximity = $max_proximity + 1; # Current minimum proximity |
| 163 |
|
| 164 |
# For each OTHER variant hitting the transcript |
| 165 |
for my $other_variant ( keys %{$aa_mutations{$transcript}} ) |
| 166 |
{ |
| 167 |
# Ignore the current mutation |
| 168 |
next if $variant eq $other_variant; |
| 169 |
# Get affected amino acids from OTHER variant |
| 170 |
my @other_affected_amino_acids = @{$aa_mutations{$transcript}{$other_variant}{mut_AAs}}; |
| 171 |
|
| 172 |
# Compare distances between amino acids |
| 173 |
my $found_close_one = 0; |
| 174 |
for my $other_variant_aa ( @other_affected_amino_acids ) |
| 175 |
{ |
| 176 |
for my $variant_aa ( @affected_amino_acids ) |
| 177 |
{ |
| 178 |
my $distance = abs($other_variant_aa - $variant_aa); |
| 179 |
# If distance is within range |
| 180 |
if( $distance <= $max_proximity ) |
| 181 |
{ |
| 182 |
$found_close_one++; |
| 183 |
$min_proximity = $distance if $distance < $min_proximity; |
| 184 |
} |
| 185 |
} |
| 186 |
} |
| 187 |
# Note that this variant is within proximity if applicable |
| 188 |
$mutations_within_proximity++ if $found_close_one; |
| 189 |
} |
| 190 |
# Now, save results in hash if there are any |
| 191 |
if ($mutations_within_proximity) |
| 192 |
{ |
| 193 |
$aa_mutations{$transcript}{$variant}{muts_within_range} = $mutations_within_proximity; |
| 194 |
$aa_mutations{$transcript}{$variant}{min_proximity} = $min_proximity; |
| 195 |
} |
| 196 |
} |
| 197 |
} |
| 198 |
|
| 199 |
# Print results |
| 200 |
my $out_fh = IO::File->new( $out_file, ">" ) or die "Couldn't open $out_file. $!"; |
| 201 |
$out_fh->print( "Mutations_Within_Proximity\tNearest_Mutation\tGene\tTranscript\tAffected_Amino_Acid(s)\tChr\tStart\tStop\tRef_Allele\tVar_Allele\tSample\n" ); |
| 202 |
for my $transcript ( keys %aa_mutations ) |
| 203 |
{ |
| 204 |
for my $variant ( keys %{$aa_mutations{$transcript}} ) |
| 205 |
{ |
| 206 |
if( exists $aa_mutations{$transcript}{$variant}{muts_within_range} ) |
| 207 |
{ |
| 208 |
my ( $gene, $chr, $start, $stop, $ref_allele, $var_allele, $sample ) = split( /\t/, $variant ); |
| 209 |
my $affected_amino_acids = join( ",", sort @{$aa_mutations{$transcript}{$variant}{mut_AAs}} ); |
| 210 |
my $line = join( "\t", $aa_mutations{$transcript}{$variant}{muts_within_range}, |
| 211 |
$aa_mutations{$transcript}{$variant}{min_proximity}, $gene, $transcript, |
| 212 |
$affected_amino_acids, $chr, $start, $stop, $ref_allele, $var_allele, $sample ); |
| 213 |
$out_fh->print( "$line\n" ); |
| 214 |
} |
| 215 |
} |
| 216 |
} |
| 217 |
$out_fh->close(); |
| 218 |
return 1; |
| 219 |
} |
| 220 |
|
| 221 |
################################################################################ |
| 222 |
|
| 223 |
=head2 get_amino_acid_pos |
| 224 |
|
| 225 |
This subroutine deducts the amino acid position within the transcript using the c_position and amino_acid_position columns in the MAF. |
| 226 |
|
| 227 |
=cut |
| 228 |
|
| 229 |
################################################################################ |
| 230 |
|
| 231 |
sub get_amino_acid_pos { |
| 232 |
# Parse arguments |
| 233 |
my $self = shift; |
| 234 |
my ( $variant_key, $mut_class, $c_position, $aa_change, $status ) = @_; |
| 235 |
|
| 236 |
# Initialize variables |
| 237 |
my $tx_start = my $tx_stop = 0; |
| 238 |
my $aa_position_start = my $aa_position_stop = 0; |
| 239 |
my $inferred_aa_start = my $inferred_aa_stop = 0; |
| 240 |
my $aa_pos = my $inferred_aa_pos = 0; |
| 241 |
|
| 242 |
# Amino acid position determination |
| 243 |
if( $aa_change && $aa_change ne "NULL" && substr( $aa_change, 0, 1 ) ne "e" ) |
| 244 |
{ |
| 245 |
$aa_pos = $aa_change; |
| 246 |
$aa_pos =~ s/[^0-9]//g; |
| 247 |
} |
| 248 |
|
| 249 |
# Parse out c_position if applicable ## |
| 250 |
if( $c_position && $c_position ne "NULL" ) |
| 251 |
{ |
| 252 |
# If multiple results, parse both ## |
| 253 |
if( $c_position =~ '_' && !( $mut_class =~ 'splice' )) |
| 254 |
{ |
| 255 |
($tx_start, $tx_stop) = split( /\_/, $c_position ); |
| 256 |
$tx_start =~ s/[^0-9]//g; |
| 257 |
$tx_stop =~ s/[^0-9]//g; |
| 258 |
|
| 259 |
if( $tx_stop < $tx_start ) |
| 260 |
{ |
| 261 |
$inferred_aa_start = $tx_stop / 3; |
| 262 |
$inferred_aa_start = sprintf( "%d", $inferred_aa_start ) + 1 if( $tx_stop % 3 ) ; |
| 263 |
$inferred_aa_stop = $tx_start / 3; |
| 264 |
$inferred_aa_stop = sprintf( "%d", $inferred_aa_stop ) + 1 if( $tx_start % 3 ); |
| 265 |
} |
| 266 |
else |
| 267 |
{ |
| 268 |
$inferred_aa_start = $tx_start / 3; |
| 269 |
$inferred_aa_start = sprintf( "%d", $inferred_aa_start ) + 1 if( $tx_start % 3 ); |
| 270 |
$inferred_aa_stop = $tx_stop / 3; |
| 271 |
$inferred_aa_stop = sprintf( "%d", $inferred_aa_stop ) + 1 if( $tx_stop % 3 ); |
| 272 |
} |
| 273 |
} |
| 274 |
else |
| 275 |
{ |
| 276 |
my ( $tx_pos ) = split( /[\+\-\_]/, $c_position ); |
| 277 |
$tx_pos =~ s/[^0-9]//g; |
| 278 |
$tx_start = $tx_stop = $tx_pos; |
| 279 |
if($tx_pos) |
| 280 |
{ |
| 281 |
$inferred_aa_pos = $tx_pos / 3; |
| 282 |
$inferred_aa_pos = sprintf( "%d", $inferred_aa_pos ) + 1 if( $tx_pos % 3 ); |
| 283 |
$inferred_aa_start = $inferred_aa_stop = $inferred_aa_pos; |
| 284 |
} |
| 285 |
} |
| 286 |
} |
| 287 |
|
| 288 |
# If we inferred aa start stop, proceed with it ## |
| 289 |
if( $inferred_aa_start && $inferred_aa_stop ) |
| 290 |
{ |
| 291 |
$aa_position_start = $inferred_aa_start; |
| 292 |
$aa_position_stop = $inferred_aa_stop; |
| 293 |
$status->{aa_position_inferred}++; |
| 294 |
} |
| 295 |
# Otherwise if we inferred aa position ## |
| 296 |
elsif( $aa_pos ) |
| 297 |
{ |
| 298 |
$aa_position_start = $aa_pos; |
| 299 |
$aa_position_stop = $aa_pos; |
| 300 |
$status->{c_position_not_available}++; |
| 301 |
} |
| 302 |
# Otherwise we were unable to infer the info ## |
| 303 |
else |
| 304 |
{ |
| 305 |
$status->{aa_position_not_found}++; |
| 306 |
$self->status_message( "Amino acid position not found for variant: $variant_key" ); |
| 307 |
return; |
| 308 |
} |
| 309 |
|
| 310 |
# Proceed if we have aa_position_start and stop ## |
| 311 |
my %mutated_aa_positions; |
| 312 |
if( $aa_position_start && $aa_position_stop ) |
| 313 |
{ |
| 314 |
for( my $this_aa_pos = $aa_position_start; $this_aa_pos <= $aa_position_stop; $this_aa_pos++ ) |
| 315 |
{ |
| 316 |
$mutated_aa_positions{$this_aa_pos}++; |
| 317 |
} |
| 318 |
} |
| 319 |
|
| 320 |
my @mutated_aa_positions = keys %mutated_aa_positions; |
| 321 |
return @mutated_aa_positions; |
| 322 |
} |
| 323 |
|
| 324 |
1; |