Entering edit mode
6.4 years ago
a01039012387
▴
10
Hello. I'm new to Varscan. I want to call the SNPs from the bam file using varscan in the order of somatic - processSomatic-fpfilter but I got an error message in the fpfilter process.
Here is my command.
perl fpfilter.pl --var-file [snp.Somatic.hc] --readcount-file [readcount] --output-file [fiinal]
and this is the error message.
Failed to interpret variant allele: V at fpfilter.pl line 333, <GEN2> line 1.
It seems to be a problem during the execution of the following command in fpfilter.
**## iupactobase - Convert IUPAC ambiguity codes to variant bases ##
sub iupactobase {
my ( $allele1, $allele2 ) = @;
return( $allele2 ) if( $allele2 eq "A" or $allele2 eq "C" or $allele2 eq "G" or $allele2 eq "T" );
# Choose the most likely base-pair, or the default for triallelic variants
if( $allele2 eq "M" ) {
return( "C" ) if( $allele1 eq "A" );
return( "A" ) if( $allele1 eq "C" );
return( "A" );
} elsif( $allele2 eq "R" ) {
return( "G" ) if( $allele1 eq "A" );
return( "A" ) if( $allele1 eq "G" );
return( "A" );
} elsif( $allele2 eq "W" ) {
return( "T" ) if( $allele1 eq "A" );
return( "A" ) if( $allele1 eq "T" );
return( "A" );
} elsif( $allele2 eq "S" ) {
return( "C" ) if( $allele1 eq "G" );
return( "G" ) if( $allele1 eq "C" );
return( "C" );
} elsif( $allele2 eq "Y" ) {
return( "C" ) if( $allele1 eq "T" );
return( "T" ) if( $allele1 eq "C" );
return( "C" );
} elsif( $allele2 eq "K" ) {
return( "G" ) if( $allele1 eq "T" );
return( "T" ) if( $allele1 eq "G" );
return( "G" );
}
die "Failed to interpret variant allele: $allele2";
}*
But the input file does not seem to be a problem. Thus, I don't know what is a problem. Of course, there is no allele V in the input file. Below is the input file for fpfilter.
**readcount**
> 1 1225707 C 27 =:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00
> A:4:36.75:30.50:31.00:4:0:0.42:0.11:187.25:4:0.01:45.00:0.01
> C:23:56.30:37.22:36.30:4:19:0.51:0.01:402.78:18:0.38:58.39:0.38
> G:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00
> T:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00
> N:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00 1 6653482 G 31
> =:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00 A:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00
> C:6:49.67:25.17:30.33:0:6:0.31:0.06:298.33:6:0.02:46.17:0.02
> G:25:58.76:29.52:36.68:22:3:0.52:0.00:564.20:25:0.64:71.88:0.64
> T:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00
> N:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00 1 9791175 T 32
> =:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00 A:14:57.79:32.21:36.43:14:0:0.64:0.03:368.86:14:0.32:55.07:0.32
> C:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00
> G:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00
> T:18:58.28:30.78:36.56:18:0:0.53:0.01:408.78:18:0.33:54.33:0.33
> N:0:0.00:0.00:0.00:0:0:0.00:0.00:0.00:0:0.00:0.00:0.00
**snp.Somatic.hc**
> chrom position ref var normal_reads1 normal_reads2 normal_var_freq
> normal_gt tumor_reads1 tumor_reads2 tumor_var_freq tumor_gt
> somatic_status variant_p_value somatic_p_value tumor_reads1_plus
> tumor_reads1_minus tumor_reads2_plus tumor_reads2_minus
> normal_reads1_plus normal_reads1_minus normal_reads2_plus
> normal_reads2_minus 1 1225707 C A 30 0 0% C 23 4 14.81% M Somatic 1
> 0.044429255 4 19 4 0 5 25 0 0 1 6653482 G C 19 0 0% G 25 6 19.35% S Somatic 1 0.046334082 22 3 0 6 14 5 0 0
As you use the standalone fpfilter, this is probably an ancient version. Get the VarScan2.4.3 java executable from Github and run
java -jar VarScan2 fpfilter
. Edit: As you say you are new to varscan, if you want you can post your entire command from the aligned reads to the final file, and we can see if you used to most recent (and recommended) settings.Oh, thanks for the reply.
Here is my entire commend for Varscan, and I already using varscan2.4.3.
and I run
java -jar VarScan.v2.4.3.source.jar fpfilter /home/varscan/snp_indel/1.snp.Somatic.hc /home/varscan/snp_indel/1.readcount --output-file /home/varscan/snp_indel/1.final --dream3-settings 1
It works! but output file doesn't have any SNPs.Sorry for coming back so late. Well what shall I say, looks like all your initial variants are junk calls, at least based on the
fpfilter
. What kind of experiment was that?