Proposed Analysis: Add CNS_region mapping in histologies file to OpenPBTA analyses

[jharenza]

What are the scientific goals of the analysis?

Add CNS_region to pbta-histologies.tsv using logic utilized on CHOP D3b end

What methods do you plan to use to accomplish the scientific goals?

Regions: Hemispheric, Midline, Spine, Ventricles, Posterior fossa, Optic pathway, Suprasellar, Other

Utilize this table to perform CNS_region mapping:

Updated CNS Region	Site	primary_site
Hemispheric	Frontal	Frontal Lobe
	Temporal	Temporal Lobe
	Parietal	Parietal Lobe
	Occipital	Occipital Lobe
	—	Right Occipital
	—	Left temporo-parietal
	—	Right Parietal
	—	Right Frontal
	—	Right Temporal
Midline	Brainstem	Brainstem
	Midbrain	Brain Stem- Midbrain/Tectum
	Pons	Brain Stem- Pons
	Medulla	Brain Stem-Medulla
	Corpus Collosum	—
	Thalamus	Thalamus
	Hypothalamus	—
	Basal Ganglia	Basal Ganglia
	Hippocampus	Hippocampus
	Pineal	Pineal Gland
	Tectum	Left Thalmus
	—	Right Thalmus
	—	Bilateral Thalmus
	—	—
Spine	Cervical	Spinal Cord- Cervical
	Thoracic	Spinal Cord- Thoracic
	Lumbar	Spinal Cord- Lumbar/Thecal Sac
	Sacral	—
		Spine
	—	Spine NOS
Ventricles	Lateral	—
	Third	—
	Fourth	—
	Cerebral aqueduct	—
	Foramen of Monro	—
	—	Ventricles
Posterior fossa	Vermis	—
	Peduncle	—
		Cerebellum
	Cerebellum	Cerebellum/Posterior Fossa
Optic pathway	Optic nerves	—
	Optic chiasm	—
	Optic tract	—
	Optic radiation	—
	—	Optic Pathway
Suprasellar	Pituitary	Suprasellar/Hypothalamic/Pituitary
	Pituitary stalk	—
Other	Meninges	Meninges/Dura
	Multifocal	—
	—	Other locations NOS
	—	Skull
	—	Cranial Nerves NOS
	—	Brain
	—	Right Occipital - Parietal Lobes and Thalamus

Reference notion doc
from @baileyckelly

What input data are required for this analysis?

pbta-histologies.tsv

How long do you expect is needed to complete the analysis? Will it be a multi-step analysis?

0.5 day

Who will complete the analysis (please add a GitHub handle here if relevant)?

@kgaonkar6

What relevant scientific literature relates to this analysis?

NA, informed by physicians: Angela Waanders and Cassie Kline

[kgaonkar6]

Is this still a to-do ticket or the base histology already takes care of this? Base histology file seems to already have the "Updated CNS Region" terms assigned.

[kgaonkar6]

The direct matches in the ticket wouldn't satisfy where primary_site values can be assigned to multiple Updated_CNS_regions, these are all "Other" in base histology.

Example list of multiple primary_site values:
primary_site

  | Cerebellum/Posterior Fossa;Temporal Lobe;Thalamus
  | Optic Pathway;Temporal Lobe
  | Optic Pathway;Suprasellar/Hypothalamic/Pituitary;Temporal Lobe
  | Cerebellum/Posterior Fossa;Spinal Cord- Thoracic
  | Basal Ganglia;Optic Pathway;Temporal Lobe;Ventricles
  | Parietal Lobe;Ventricles
  | Skull;Temporal Lobe
  | Cerebellum/Posterior Fossa;Spinal Cord- Cervical;Spinal Cord- Lumbar/Thecal Sac;Spinal Cord- Thoracic
  | Cerebellum/Posterior Fossa;Frontal Lobe
  | Brain Stem- Pons;Cerebellum/Posterior Fossa

For these, let's annotate as "Mixed", knowing they need pathology review from Cassie for a primary site region designation.

[jharenza]

Is this still a to-do ticket or the base histology already takes care of this? Base histology file seems to already have the "Updated CNS Region" terms assigned.

Converting all of the extraction of data using DBT will mean that CNS_region will no longer be generated in the base histology file from the D3b end. However, we can add this, as you suggest, to our base histology preparation following QC.

[jharenza]

Closing, as this is now a part of the intermediate D3b workflow prior to release of pbta-histologies-base.tsv per this comment.

[jharenza]

@kgaonkar6 I am going to re-open this, as I think the logic was not captured entirely as anticipated for #849. For instance, some samples which should be mixed are annotated as midline or hemispheric. We can update this in #v19.

[jharenza]

One other thing to note is that for our "Mixed" samples, @jainpayal022 and Cassie Kline will be manually reviewing pathology and imaging reports of about 70 or so HGAT samples to assess the primary site of origin. We will then add these manually curated values for those CNS_regions. See this ticket.

[kgaonkar6]

Adding the code here for CNS_region matching here https://github.com/d3b-center/D3b-codes/blob/master/OpenPBTA_v19_release_QC/QC_histology_v19.Rmd to record

[kgaonkar6]

Just wanted to add the primary site changes from the latest pulls, so guide updates to the above primary_site ~ CNS_region matches

Kids_First_Biospecimen_ID	primary_site_latest	primary_site_previous
BS_1Q524P3B	L. Pons Anterior	Pons/Brainstem
BS_22VCR7DF	L. Lateral Pons	Pons/Brainstem
BS_5968GBGT	R. Posterior Pons; Adjacent #6	Pons/Brainstem
BS_AF5D41PD	L. Frontal Periventricular White Matter; Adjacent #3	Pons/Brainstem
BS_AK9BV52G	Cerebellar White Matter Adjacent Necrosis + Medulla	Pons/Brainstem
BS_D6STCMQS	L. Anterior Medulla	Pons/Brainstem
BS_EE73VE7V	R. Inferior Pons	Pons/Brainstem
BS_HYKV2TH9	R. Anterior Pons; Adjacent #7	Pons/Brainstem
BS_J8EH1N7V	Inferior Pons	Pons/Brainstem
BS_J8EK6RNF	Brain Stem-Medulla;Brain Stem- Midbrain/Tectum;Brain Stem- Pons;Cerebellum/Posterior Fossa;Thalamus	Pons/Brainstem
BS_X5VN0FW0	Inferior Medulla	Pons/Brainstem
BS_Y74XAFJX	Superior Pons	Pons/Brainstem
BS_YHXMYDBN	L. Pons	Pons/Brainstem

[kgaonkar6]

I'm not sure where exactly we will be using CNS_regions so just wondering if we have an updated dictionary matching primary_site to CNS_region to incorporate the new primary_sites assigned to PNOC autopsy samples? Currently the CNS_region is being assigned as NA since it doesn't match any terms in the issue description.

[jharenza]

Thank you for checking on this @kgaonkar6 - I had the answers from Cassie, but never posted here. See below with the only one in question being Cerebellar White Matter Adjacent Necrosis + Medulla, so for now, we can designate as Mixed, unless there is other data available to confirm Midline:

Kids_First_Biospecimen_ID	primary_site_latest	primary_site_previous	CNS_region
BS_1Q524P3B	L. Pons Anterior	Pons/Brainstem	Midline
BS_22VCR7DF	L. Lateral Pons	Pons/Brainstem	Midline
BS_5968GBGT	R. Posterior Pons; Adjacent #6	Pons/Brainstem	Midline
BS_AF5D41PD	L. Frontal Periventricular White Matter; Adjacent #3	Pons/Brainstem	Mixed
BS_AK9BV52G	Cerebellar White Matter Adjacent Necrosis + Medulla	Pons/Brainstem	Mixed
BS_D6STCMQS	L. Anterior Medulla	Pons/Brainstem	Midline
BS_EE73VE7V	R. Inferior Pons	Pons/Brainstem	Midline
BS_HYKV2TH9	R. Anterior Pons; Adjacent #7	Pons/Brainstem	Midline
BS_J8EH1N7V	Inferior Pons	Pons/Brainstem	Midline
BS_J8EK6RNF	Brain Stem-Medulla;Brain Stem- Midbrain/Tectum;Brain Stem- Pons;Cerebellum/Posterior Fossa;Thalamus	Pons/Brainstem	Midline
BS_X5VN0FW0	Inferior Medulla	Pons/Brainstem	Midline
BS_Y74XAFJX	Superior Pons	Pons/Brainstem	Midline
BS_YHXMYDBN	L. Pons	Pons/Brainstem	Midline

[kgaonkar6]

Thank you for the update!

[jharenza]

Confirmed this will remain Mixed

Cerebellar White Matter Adjacent Necrosis + Medulla, so for now, we can designate as Mixed

[kgaonkar6]

Previously the first step of CNS_region assignment is "Mixed" is there are multiple values separated by ";" and then check if all individual values are part of on CNS_region. So the primary_site like R. Posterior Pons; Adjacent #6,R. Anterior Pons; Adjacent #7 and Brain Stem-Medulla;Brain Stem- Midbrain/Tectum;Brain Stem- Pons;Cerebellum/Posterior Fossa;Thalamus will be assigned Mixed.

https://github.com/d3b-center/D3b-codes/blob/20210212-release/OpenPBTA_v20_release_QC/code/util/primary_site_matched_CNS_region.R do we want to update the code or make these exeptions since "Adjacent" is not a brain location per se?

[jharenza]

To be honest, I was thinking about this and think that whoever ingests this into the BRP should match these regions with CBTN terms so we don't always have to update the terms. Let me check in with Jenn Mason and Shannon Robbins.

In the meantime. I was envisioning these going into the JSON file for Midline.

[kgaonkar6]

Just need a confirmation

BS_J8EK6RNF seems to have primary_site "Pons" in the latest histology file pulls ( since 20210126-data release ) instead of the "Brain Stem-Medulla;Brain Stem- Midbrain/Tectum;Brain Stem- Pons;Cerebellum/Posterior Fossa;Thalamus" in the table above do we need to confirm this @jharenza ?

[jharenza]

Just need a confirmation

BS_J8EK6RNF seems to have primary_site "Pons" in the latest histology file pulls ( since 20210126-data release ) instead of the "Brain Stem-Medulla;Brain Stem- Midbrain/Tectum;Brain Stem- Pons;Cerebellum/Posterior Fossa;Thalamus" in the table above do we need to confirm this @jharenza ?

Let me check on this. Are the other samples matching primary_site_latest?

[kgaonkar6]

yes

[jharenza]

Ok, it looks like Brain Stem-Medulla;Brain Stem- Midbrain/Tectum;Brain Stem- Pons;Cerebellum/Posterior Fossa;Thalamus was never a value in @jainpayal022's excel sheet for kids first import, but rather, Pons, so waiting for her to confirm this longer string was a mistake along the way somehow.

[jharenza]

confirmed that this value should be Pons

[jharenza]

we added this upstream during QC