Add additional bulk upload options for PhenoTips fields

[AliciaByrne]

Would be helpful to have bulk upload options for PhenoTips fields in addition to 'HPO Terms present' and 'HPO Terms absent'. Key fields of interest (for CMG_Scott projects) are: 'Candidate Genes', 'Previously Tested Genes', and 'Diagnosis' (as below). Would also be helpful to bulk upload 'maternal/paternal ethnicity' and 'consanguinity' (below) to cross-check with inferred results from sequencing.

Scott project data is exported from REDCap to an Excel file. Some fields (eg HPO terms) are currently split over multiple cells - we can condense these to single cell, comma separated. There is also a free-text phenotype entry field with additional info not necessarily matching to HPO term - unsure where this would best fit in Seqr.

[hanars]

Hi Alicia,

When looking over the sample file you provided, I encountered the following concerns

• Many of the "Diagnosis" currently contain commas, so condensing them into a single comma-separated column would cause problems. I would recommend that for all the fields that need to be condensed into a single column you use ";" as the new separator
• The diagnosis section in phenotips takes a list of valid omim disorders and a free text field. However, the diagnosis included in the sample you provided me do not map perfectly to OMIM. For example, one of the diagnoses provided was "Cornelia de Lange syndrome", but in OMIM there are 5 different "Cornelia de Lange syndrome"s so I wouldn't know which one to use. As an even harder example, another provided diagnosis is "Short rib-polydactyly syndrome, Beemer-Langer type" which does not neatly map onto any entry in omim. There is a "SHORT-RIB THORACIC DYSPLASIA 12; SRTD12" which includes under alternative titles "SHORT RIB-POLYDACTYLY SYNDROME, TYPE IV; SRPS4 SRPS IV", "BEEMER-LANGER SYNDROME" and "SHORT RIB SYNDROME, BEEMER TYPE". However, there is no reliable automated way to figure out that that is likely the omim condition in question. The possible solutions for this that I can see are
  • before uploading, someone manually ensures that the diagnoses all map to valid omim entries
  • add all the diagnoses to the "additional comments" section instead of the "omim disorder" list
  • if a diagnoses does map to an omim entry add it as an "omim disprder", otherwise add it to the "additional comments"
• the previously tested genes section is a list of gene-comment pairs. However, the columns you have labeled as "previously tested genes" are actually the types of tests that were done, i.e. "Panel - Cornelia de Lange syndrome". We could add a notes field under "Medical History" and put this information there instead, or we could still add it to "previously tested genes" but with the gene part left blank
• for the candidate genes, some are actually just a gene name while others look like "DKN1C (esp PCNA binding domain) IMAGE syndrome SAMD9- Mirage syndrome (probably should do SNP array at some stage )". Since this should ultimately be a list of gene-comment pairs it would be helpful if it could be formatted in a way that could be parsed such as "DKN1C - (esp PCNA binding domain) IMAGE syndrome; SAMD9- Mirage syndrome (probably should do SNP array at some stage )" with ";" separating the genes and "-" separating the gene name from the comment
• for the free text phenotypes, I think that could go under "Indication for referral" in the "Patient information" section, or we could add a notes field under "Medical History"

Let me know what you think

[AliciaByrne]

Hi Hana

Thanks for your suggestions! I've responded to them below in the same order as you've raised them.

1. Happy for all fields that need to be condensed into a single column to use ";" as the new separator.
2. Unfortunately we're not able to integrate omim disorders into our internal database so we've been using ORDO nomenclature which you're right, does not map perfectly to omim. Let's put all the diagnoses to be added into the "additional comments" section instead of the "omim disorder" list.
3. I think it would be good to keep the information about previous testing under "previously tested genes", so let's add the info to the comments section with the gene part left blank.
4. Happy to format candidate gene lists with ";" separating the genes and "-" separating the gene name from the comment. Would it be helpful to do this for 'previously tested genes' too in cases where there is a gene name listed?
5. Let's put the free text phenotypes under "Indication for referral" in the "Patient information" section, that way it's up the top and easily accessible when you open the Phenotips doc - most information here should be translated to HPO terms under major or minor phenotype but there are some that don't map exactly or aren't covered.

Thanks! :)

[hanars]

Re: your question on 4 "Would it be helpful to do this for 'previously tested genes' too in cases where there is a gene name listed?" I would say yes thats a good idea. If there isn't a gene for that field you could just start the line with a "-" (i.e. "- Panel" instead of just "Panel")

If you could create a new test file based on the changes discussed here and send it to me that would be really helpful. Slacking it to me the way you did the first one works fine

[AliciaByrne]

Sorry for the delay getting to this. Have just slacked you the new test file.

Also - we've decided to take out the free text phenotype information, as often times the clinicians will just copy and paste the autopsy report and there is sometimes identifying information. The majority of this information is captured in the HPO terms and diagnosis sections so we will not lose too much information by excluding it.

[hanars]

Hi, I'm sorry for the massive delay on this. As you may have realized, seqr no longer integrates directly with phenotips. Are these still data fields you would like to bulk upload directly to seqr?

[hanars]

Based on the new fields available in seqr now that we don't use phenotips, I think we should add the following fields as options for bulk upload. They should be added as optional to the EditHPOBulkForm, and we should rename that form to something more descriptive (maybe "Edit Phenotypes")

• candidate_genes
• rejected_genes
• disorders (validate valid OMIM number)
• maternal_ethnicity
• paternal_ethnicity
• consanguinity

[hanars]

Additional fields requested in https://github.com/broadinstitute/seqr/issues/1759 (this new field list includes everything previously required by this ticket)

• Age
• Age of onset
• Consanguinity
• Other Affected Relatives
• Expected Mode of Inheritance
• Assisted Reproduction
• Maternal Ancestry
• Paternal Ancestry
• Pre-discovery OMIM disorders
• Previously Tested Genes
• Candidate Genes
• Case Review Status

[hanars]

Note: Case Review Status is a protected field so bulk updating it will happen in the main "update individual" form not the "update individual metadata" form

[hanars]

the "update individuals" bulk upload now supports all of these fields, including as json exported from a separate phenotips instance