search

konradjk / loftee

MIT License
172 stars 55 forks source link

LoF_flags are not being properly annotated #26

Open Madelinehazel opened 6 years ago

Madelinehazel commented 6 years ago

Hello,

I have noticed that when running LOFTEE, the LoF_flags are not being properly annotated; they are always blank in the output. So, for instance, rs201677741 from the ExAC vcf is annotated by our installation of LOFTEE as follows:

14_24709750_G/T 14:24709750 T 26277 NM_001099274.1 Transcript stop_gained 1278/1852 936/1356 312/451 Y/* taC/taA rs201677741 LoF_flags=;IMPACT=HIGH;LoF_filter=;LoF=HC;LoF_info=POSITION:0.690265486725664,PHYLOCSF_TOO_SHORT;STRAND=-1;VARIANT_CLASS=SNV;SYMBOL=TINF2;BIOTYPE=protein_coding;CANONICAL=YES;ENSP=NP_001092744.1;EXON=6/9;HGVSc=NM_001099274.1:c.936C>A;HGVSp=NP_001092744.1:p.Tyr312Ter;GMAF=T:0.0002;AFR_MAF=T:0;AMR_MAF=T:0;EAS_MAF=T:0;EUR_MAF=T:0.001;SAS_MAF=T:0;AA_MAF=T:0.0002;EA_MAF=T:0.0001;ExAC_MAF=T:3.720e-04;ExAC_Adj_MAF=T:0.0003727;ExAC_AFR_MAF=T:0.0002041;ExAC_AMR_MAF=T:0;ExAC_EAS_MAF=T:0;ExAC_FIN_MAF=T:0;ExAC_NFE_MAF=T:0.0006444;ExAC_OTH_MAF=T:0;ExAC_SAS_MAF=T:0

Ás you can see, there is no flag after LoF_flags (bolded). However, if I look at this variant in the gnomAD browser (see attached image), it is given the flag PHYLOCSF_WEAK.

loftee

We have run a few variants and we have not seen LoF_flags where we should, even though they are annotated as such in gnomAD.

This is our code: perl /opt/tools/ensembl/ensembl-tools/scripts/variant_effect_predictor/variant_effect_predictor.pl \ --cache --dir_cache=/opt/tools/ensembl/cache/ \ -i dummy.vcf \ -o /mnt/causes-data01/new/COUSE/test/dummy.out \ --plugin LoF,loftee_path:/opt/tools/loftee,human_ancestor_fa:/opt/tools/loftee/human_ancestor.fa,check_complete_cds,conservation_file:mysql \ --dir_plugins=/opt/tools/loftee \ --everything \ --assembly GRCh37 \ --port 3337 \ --force \ --refseq \ --total_length \ --fasta /opt/tools/ensembl/tools_data/vep/homo_sapiens/Homo_sapiens.GRCh37.dna.toplevel.fa We have tried pointing to the downloaded conservation_file, phylocsf.sql, and alternatively, loading the source file in MySQL.

Help with this issue would be greatly appreciated!

Thanks, Maddie

bryketos commented 6 years ago

Thanks for the post Maddie - looks like great work on your end and you have a nice annotation of a deleterious variant. LoF_flags might be empty if no flags are explicitly instantiated by the plugin, e.g. a NAGNAG or intronless pre-mRNAs. Are you expecting LoF_flags to always have a value? Thanks, Brian

bryketos commented 6 years ago

@konradjk Feel free to comment re: gnomad. OT: great progress on this package - I made a Perl version of MaxEntScan that uses data submodules which might make a nice PR - feel free to check it out on CPAN at http://search.cpan.org/~bryketos/Bio-RNA-SpliceSites-Scoring-MaxEntScan-0.05/lib/Bio/RNA/SpliceSites/Scoring/MaxEntScan.pm or on github at https://github.com/bryketos/Bio-RNA-SpliceSites-Scoring-MaxEntScan. Cheers!

Madelinehazel commented 6 years ago

Hi Brian,

Thanks for replying. No, I don't expect LoF_flags to always have a value, but we have tested a number of variants that are present in gnomAD that do have LoF_flags, and the output does not contain the LoF_flag. Unfortunately, there is no error message, so I am having trouble diagnosing what the problem is.

rvernica commented 5 years ago

@Madelinehazel where you able to find out why LoF_flags are empty?

konradjk commented 5 years ago

Apologies for the delay - preparing a more final release, so wanted to check in: are you still having this problem? And is anything printed to the console that might look worrying?