[MAF/VCF] `pymaf.MafFrame.from_vcf` assumes CSQ is the first field in the INFO record about fuc

[lbeltrame]

It looks like it's working on my test VCF (~1000 variants annotated with VEP):

In [1]: from fuc import pyvcf, pymaf

In [2]: mf = pymaf.MafFrame.from_vcf("test_variants.vcf.gz")

In [3]: mf
Out[3]: <fuc.api.pymaf.MafFrame at 0x129285a90>

In [2]: mf = pymaf.MafFrame.from_vcf("test_variants.vcf.gz")

In [3]: mf
Out[3]: <fuc.api.pymaf.MafFrame at 0x129285a90>

In [4]: mf.df
Out[4]:
     Hugo_Symbol   Entrez_Gene_Id Center  ... Tumor_Seq_Allele2 Protein_Change  Tumor_Sample_Barcode
0           MTOR  ENSG00000198793      .  ...                 A              .                   117

[sbslee]

@lbeltrame,

You are absolutely correct here. Thanks for reporting the problem. I fixed this issue in the 0.34.0-dev branch as per your suggestion (i.e. split by ; and then look for CSQ). See an example below (note that the first row uses the line you provided in the original post):

>>> from fuc import pyvcf, pymaf
>>> data = {
...     'CHROM': ['chr1', 'chr2'],
...     'POS': [100, 101],
...     'ID': ['.', '.'],
...     'REF': ['G', 'T'],
...     'ALT': ['A', 'C'],
...     'QUAL': ['.', '.'],
...     'FILTER': ['.', '.'],
...     'INFO': ['AC=2;ACGTNacgtnMINUS=0,0,0,0,0,0,0,0,0,0;ACGTNacgtnPLUS=5,0,61,0,0,0,0,0,0,0;AN=4;AS_FilterStatus=SITE;AS_SB_TABLE=9,41|1,5;CALLERS=mutect2;CSQ=A|3_prime_UTR_variant|MODIFIER|MTOR|ENSG00000198793|Transcript|ENST00000361445|protein_coding|58/58||ENST00000361445.9:c.*700C>T||8471/8721|||||||-1|||SNV|HGNC|HGNC:3942|YES|1|P1|CCDS127.1|ENSP00000354558|P42345||UPI000012ABD3||Ensembl|G|G||1|||||chr1:g.11106785G>A|||||||||||||||||||||||||||;ClippingRankSum=-0.79;DKFZBias=damage;DP=85;ECNT=2;EPR=pass;FS=0;GERMQ=93;MBQ=34,33;MFRL=153,154;MMQ=60,60;MPOS=28;MQ=60;MQ0=0;MQRankSum=0;POPAF=7.3;ReadPosRankSum=0.452;TLOD=12.45', 'AC=9;CSQ=C|splice_donor_variant|HIGH|MTOR|2475|Transcript|NM_001386500.1|protein_coding||46/57||||||||||-1||EntrezGene||||||||A|A|||||||||||||||||||||||||||||'],
...     'FORMAT': ['GT:AD:DP:AF', 'GT:AD:DP:AF'],
...     'A': ['0/1:176,37:213:0.174', '0/1:966,98:1064:0.092']
... }
>>> vf = pyvcf.VcfFrame.from_dict([], data)
>>> vf.df
  CHROM  POS ID REF  ... FILTER                                               INFO       FORMAT                      A
0  chr1  100  .   G  ...      .  AC=2;ACGTNacgtnMINUS=0,0,0,0,0,0,0,0,0,0;ACGTN...  GT:AD:DP:AF   0/1:176,37:213:0.174
1  chr2  101  .   T  ...      .  AC=9;CSQ=C|splice_donor_variant|HIGH|MTOR|2475...  GT:AD:DP:AF  0/1:966,98:1064:0.092

[2 rows x 10 columns]
>>> mf = pymaf.MafFrame.from_vcf(vf, keys=['AD', 'AF'])
>>> mf.df
  Hugo_Symbol   Entrez_Gene_Id Center NCBI_Build  ... Protein_Change  Tumor_Sample_Barcode      AD     AF
0        MTOR  ENSG00000198793      .          .  ...              .                     A  176,37  0.174
1        MTOR             2475      .          .  ...              .                     A  966,98  0.092

[2 rows x 17 columns]
>>> 

You can install the development branch:

$ git clone https://github.com/sbslee/fuc
$ cd fuc
$ git checkout 0.34.0-dev
$ pip install -e .

Please let me know if this doesn't solve the issue.

[sbslee]

Great to hear! Thanks again for reporting. The official release for 0.34.0 will be made some time next month. Please feel free to reopen this issue if necessary.