Is "Sex" correctly labelled in the Pedigree example and Sex documentation

[slaurie]

Hi @julesjacobsen

Is Sex correctly labelled in the Pedigree/Sex schemas?

In genetics/genomics, we usually use "1" for male, and "2" for female (easy to remember as equivalent to the normal number of X chromosomes per sex), but in the documentation and the examples for the Pedigree (https://phenopacket-schema.readthedocs.io/en/latest/pedigree.html) and Sex fields (https://phenopacket-schema.readthedocs.io/en/latest/sex.html#rstsex), it indicates that females should be encoded as 1 and males as 2.

Unfortunately the NICT links on the Sex page didn't work for me, so I can't check if they do have F1 and M2. However, if this is the case, then there is an inconsistency between NICT and what is defined in PLINK, so a choice needs to be made, as the documentation is inconsistent. Personally, I would suggest we stick with PLINK.

This may seem trivial, but obviously could lead to a lot of miscoding that will be a pain for users to rectify later.

[slaurie]

P.S. If we were to stick with PLINK values, this would also affect Karyotipic Sex values
https://phenopacket-schema.readthedocs.io/en/latest/karyotypicsex.html#data-model

[julesjacobsen]

Yes, they are correct - note that the ORDINAL value isn't what should be used in the phenopacket, the value in the NAME field for these docs is what should be used, as shown in the example : https://phenopacket-schema.readthedocs.io/en/latest/sex.html#example

Example code for converting from PED /PLINK to Phenopackets would be

int pedSex = ...
var phenopacketSex = switch(pedSex) {
    case 1 -> FEMALE;
    case 2 -> MALE;
    case 3 -> OTHER_SEX;
    default -> UNKNOWN_SEX;

It's unfortunate that the ordinals don't match the PLINK values but they are different things.

[slaurie]

Thanks for confirming, that your encoding is correct, and I noticed that the actual terms MALE and FEMALE are used explicitly which is great.

However, there is still an error on the "Pedigree" documentation (https://phenopacket-schema.readthedocs.io/en/latest/pedigree.html),
where it states:

"In a PED file, the sex of individuals is encoded as a “1” for females, “2” for males, and “0” for unknown. Phenopackets uses Sex instead.",

and this is also the encoding you have used in the examples., whereas the PLINK PED documentation (https://zzz.bwh.harvard.edu/plink/data.shtml) has

"Sex (1=male; 2=female; other=unknown)"

So it might be worth correcting this somehow. I do much prefer the explicit MALE and FEMALE.

Thanks, Steve

[pnrobinson]

@slaurie thanks for this.
I am update the documentation as follows:

In a PED file, affected persons are encoded with "2", and unaffecteds by "1"
(a "0" is used if no information is available). Instead, Phenopackets uses an enumeration as shown in the table.

In a PED file, the sex of individuals is encoded as a "2" for females, "1" for males, and "0" for unknown. Phenopackets
uses :ref:`rstsex` instead (Note that the enumeration used for Sex by the Phenoapcket Schema has 1 for FEMALE and 2 for MALE, but this is internal and is not visible in the JSON-serialized Phenopacket).