completing descriptions and validation for genomic alteration annotations #28
Conversation
There was a problem hiding this comment.
Hi @aim11,
Thanks for updating the validations with the descriptions and expected values! In some cases I still got some validation errors with the new data, so I put them in the comments for you to double-check.
Another thing we should maybe add in the validation files is a strict: true clause just under validate:, so that an error is raised in case some columns are not accounted for by the validation scripts, but present in the data.
| nullable: true | ||
| checks: | ||
| str_matches: ^[0-9]{7,}(,[0-9]{7,})*$ | ||
| str_matches: ^[0-9]{7,}(;[0-9]{7,})*$ |
There was a problem hiding this comment.
On my side I'm getting an error here because of the semicolon? But the version with the colon ( ^[0-9]{7,}(,[0-9]{7,})*$ ) is valid.
| - Likely Oncogenic | ||
| - Oncogenic | ||
| - Unknown | ||
| - known # Not present in source documentation description. |
There was a problem hiding this comment.
Here known still comes out as a value present in the data.
| dtype: str | ||
| description: "Describes the therapeutic implication that applies to the indication in cancer" | ||
| nullable: true | ||
| referenceGenome: |
There was a problem hiding this comment.
I get an error here stating that referenceGenome is not present in the data frame.
| # nullable: true # FIXME: Encountered point '.' values in data. | ||
| description: "ClinVar variant status" | ||
| nullable: true | ||
| clinvar_assoc: |
There was a problem hiding this comment.
I also get an error here stating clinvar_assoc is not present in the data frame.
| # nullable: true # FIXME: Encountered point '.' values in data. | ||
| description: "ClinVar databse ID" | ||
| nullable: true | ||
| clinvar_sig: |
There was a problem hiding this comment.
Same here: clinvar_sig not present in data frame.
| dtype: int64 # FIXME defined as 'float' in documentation. | ||
| dtype: float64 | ||
| description: "Lower bound for expected homogeneity confidence interval" | ||
| hom_hi: |
There was a problem hiding this comment.
For hom_hi, I still get the type encountered is int64.
| value: | ||
| - HET | ||
| - LOH | ||
| hom_lo: |
There was a problem hiding this comment.
Same as with hom_hi, for hom_lo I still get the encountered type is int64.
| hom_pbinom_lo: | ||
| dtype: float64 | ||
| description: "Binomial distribution probability for expected homogeneity" | ||
| homogenous: |
There was a problem hiding this comment.
Here I still get that the type encountered is object for homogenous and not boolean.
| nMinor: | ||
| dtype: int64 | ||
| description: "Minor allele copynumber (from CNAs if available)" | ||
| nMajor: |
There was a problem hiding this comment.
I still get that there are some nullable values here for nMajor, and also that the values encountered are float64.
| dtype: float64 | ||
| description: "Protein change by RefGene" | ||
| nullable: true | ||
| nMinor: |
There was a problem hiding this comment.
For nMinor I also get that there are nullable values, and the type is a float64.
#27