How to link DNA data with RNA data in Scout

[Vince-janv]

Background

From a clinician's perspective RNA data complements the DNA data. Therefore having them both available in the same "Case" view in Scout is of great value. Here's an example in scout stage.

The current mapping for uploading a case (cg only works with case-wise uploads) goes from an RNA-case via it's samples. For each sample it finds all samples within the same customer collaboration that fulfil the criteria below:

• Have the same subject_id
• Have the same value for is_tumour (true/false)
• Is a DNA sample (by checking prep-categories)

The rough idea is sketched out below

graph LR
    RNA_trio --> RNA_child
    RNA_trio --> RNA_mother
    RNA_trio --> RNA_father


    RNA_child --> DNA_child_muscle
    RNA_child --> DNA_child_blood
    RNA_father --> DNA_father
    RNA_mother --> DNA_mother

    DNA_child_blood --> DNA_singleton
    DNA_child_blood --> DNA_trio

    DNA_father --> DNA_trio

    DNA_mother --> DNA_trio

    DNA_child_muscle --> DNA_singleton_2

subgraph "Cases to upload to"
	DNA_singleton
	DNA_trio
	DNA_singleton_2
    end

Loading

Problem

This approach causes a lot of ambiguity, since multiple dna samples can fulfil these criteria, and a sample can be part of multiple cases. This results in a situation where when asking which case a RNA-case is linked to, you get a list of cases back. Currently the data is then uploaded to all these cases.

Appart from general confusion this approach cannot resolve between samples of multiple sources. One would for example not want to upload RNA-results from a blood sample, to a DNA-sample from the actual tumour.

There are currently 4 issues in cg stemming from this logic and more are being created regularly indicating that the logic is very brittle.

• Subject_ID not enough to distinguish samples #1600
• mip-rna Scout upload fails if there is a cancelled dna sample for the same subjectID #2082
• Automatic re-upload of existing rna-analysis to "primary analyses" #1977
• Bug in upload of RNA cases to Scout #1870

What is needed

Since the customers placing orders know which samples/cases to link we want them to explicitly state so. This needs to be done in a way where the customers do not have to perform a lot of clicks in the orderportal. Ideally it should be available easily via JSON, excel as well.

[Vince-janv]

Ideas brought up during meeting 2024-08-29

Explicit on case

If the customer wants RNA-data linked to a scout case they need to explicitly provide which case to link to. subject_id will still be needed to match samples within a case.

Pros

• Removes almost all of the complex logic in cg

Cons

• Data will only be uploaded to one case and not all cases containing containing a corresponding DNA-sample
• One more field (per case) for customers to fill out
• Future DNA analyses will not contain the RNA-data

---

Explicit on samples

If the customer wants RNA-data linked, they explicitly match the RNA-samples to which DNA-samples they correspond. The provided subject_ids should be used fetch suggestions.

Pros

• Removes half of the complex logic for uploading
• RNA data will be uploaded to all previous DNA-cases with any of the selected samples in them

Cons

• We will still have logic where one case is uploaded to multiple cases
• We will add one more field (per sample) for the customer to fill in
• Future DNA analyses will not contain the RNA-data

---

Add new database hierarchy

A scout "Case" represents a case in the medical sense. A patient or group to treat. At CG Case is synonymous to an analysis. Ie a set of samples in and results returned. For a medical case one can perform multiple analyses (DNA, RNA etc.) which creates a mismatch between what a scout-case and cg-case is.

To solve this one could make create a "real" case in cg (referred to as meta-case in the meeting) to which different kinds of analyses (what we now call case) can be added.

Adding something like this would require a lot of planning before implementing, since 90% of the cg-logic revolves around the case-model.

Pros

• More future proof and expandable for future applications

Cons

• Requires a large re-write of cg-logic
• Not sure how translatable this is to non-human samples

[Vince-janv]

Further discussion during meeting 2024-09-09

The meeting focused a lot on envisioning what scout is and how it is used by the clinicians. The following thoughts were put forward:

Vision

• A scout case contains individuals who in practice can have multiple samples each.
• Data from different samples from the same individuals should be uploaded to the same scout case. IE DNA and RNA samples contained within the same scout case.
• A cg case just contains which samples should be run in which pipelines.
• The users should have the option to upload results from multiple cg cases containing the same individuals. Meaning DNA and RNA (and maybe other types if data) should be uploaded to the same scout case.
• It should not matter what type of analysis is ordered first. If no DNA data exists in scout, an RNA scout case should be able to be created and DNA data uploaded to it later if needed

Discussed Implementation

CG

• Details regarding scout upload should be kept in a new table somehow linked to the data_delivery for the case.
• The information regarding where to upload the scout case needs to be provided by the customers during the ordering.
• Since we still need to map samples together within the case, subject_id is still needed

Orderportal

• If scout upload is selected the user should have the option to select an existing case to link it to.
  • How do we get a list of "scout cases"?

Scout

• In the Load config one should be able to list multiple samples for the same individual
  • How is the individual identified? Should we use subject_id in the load config?

Short term solution

Since the suggestions above require big changes in multiple systems an interim solution is needed. We agreed on the following.

• Keep the logic used for MIP-RNA
• Extract a list of subject_ids from the clinicians LIMS and populate the cg-database with them