Making PHESANT compatible dataset from participant table

Former User of DNAx Community_65
Hi   I want to use PHESANT:   https://ukbiobank.dnanexus.com/app/phesant   The usage states that the phenotype file must contains columns such as x[varid]_[instance]_[array].   However, participant tables embed the array data into a single column:   https://dnanexus.gitbook.io/uk-biobank-rap/working-on-the-research-analysis-platform/using-spark-to-analyze-tabular-data#database-columns   Is there code available to convert data from a participant table to a format compatible for PHESANT, ie splitting the embedded array to multiple columns?   Thanks,   Vince    

Comments

2 comments

Date Votes
  • Comment author
    Ondrej Klempir DNAnexus Team

    Table exporter (https://ukbiobank.dnanexus.com/app/table-exporter) has an option for UKB-style headers.

     

    In addition, you may consider writing a function which converts the UKB trait IDs to the format 'x[varid]_[instance]_[array]' required by PHESANT. It should not be too complicated.

     

    # i.e:

    # p31 -> x31_0_0

    0
  • Comment author
    Former User of DNAx Community_65

    Thanks for the reply. It's unclear whether table exporter will split the embedded array. For example HES from p41270 are within a single embedded array that contain multiple ICD10 codes. In the native UKB format these are present as separate columns.

    0

Please sign in to leave a comment.