arctos
arctos copied to clipboard
ArctosDB
disposition cleanup: subsampled
Ref: https://github.com/ArctosDB/arctos/issues/2667
Problem: Value makes no sense.
Proposed solution:
- update disposition="subsampled" to disposition="in collection"
- append "former disposition: subsampled" to part remarks
I will proceed with this if there are no objections by ~2022-06-17.
Users:
-------------+-------
UCM:Mamm | 8
ASNHC:Mamm | 8
DMNS:Bird | 3
DMNS:Para | 1
UAM:Fish | 35
UAM:Herb | 2
UTEP:Bird | 1
DMNS:Mamm | 1303
MSB:Mamm | 3
UWBM:Mamm | 3
UAMObs:Ento | 2
UNCG:Para | 4
MSB:Herp | 2
UAM:ES | 1
MSB:Host | 18
UCM:Bird | 1
BYU:Herp | 5
UMNH:Mamm | 31
MSB:Bird | 3
UAM:Ento | 4
UCM:Herp | 4
UTEP:ES | 7
MVZ:Bird | 41
MSB:Para | 25
Contacts:
@ebraker @bryansmclean @mkoo @campmlc @ccicero @dssikes @mvzhuang @jtgiermakowski @jebrad @AdrienneRaniszewski @acdoll @jldunnum @jrdemboski @byuherpetology @msbparasites @mlbowser @kderieg322079 @StefanieBond @catherpes,@catherpes @SerinaBrady
Data:
Wouldn't it also be appropriate to add a subsample part to all of these?
add a subsample part to all of these
Some/most of them are derivatives.
part_name | has_parent
------------------------------------+------------
skeleton | yes
whole organism | no
heart, kidney, liver, spleen | yes
tissue | no
heart, liver | yes
heart, liver, muscle | yes
skin clip | no
liver | yes
heart | no
kidney | yes
fin | yes
tissue | yes
skin, study | yes
skeleton | no
feces | yes
whole organism | yes
unknown | yes
radius | no
skin, study | no
heart | yes
stomach | no
heart, kidney, muscle | yes
toe | yes
tooth | yes
heart, kidney | yes
carcass | yes
tibia | yes
skin | yes
radius | yes
kidney, muscle | yes
skull | yes
spleen | no
heart, spleen | yes
heart, kidney, liver, lung, spleen | yes
heart, kidney, liver | yes
dentary | yes
heart, kidney, liver, muscle | yes
observation | no
liver | no
mandible | yes
kidney, liver | yes
skull | no
lung | yes
femur | no
metacarpal | yes
muscle | yes
I really like the "subsampled" value because when an outside user sees "whole organism" in the part table and then "whole organism" a row below slightly indented, the "subsampled" status helps clarify what is going on. Also, many subsamples aren't in fact "in collection" but rather loaned/consumed, so regardless, the proposed change to update disposition to "in collection" may not be suitable.
I've been thinking for some time that it would be great if we had some means of automatically indicating when a part had been subsampled, when we check "subsample y/n?" on the loan forms. I subsample virtually everything, and I agree that the little indentation of the subsample is not obvious to anyone not doing this on a daily basis. But I haven't been changing the disposition to subsample - my subsamples are are consumable with disposition"on loan" . Maybe instead we could get a choice on the loan page to select the new part name (not currently an option - you get an indented "whole organism" or "heart, kidney, liver, spleen"). Some other way to indicate parent/child relationships? We have different part IDs, with indication of the parent part ID visible curatorially on Edit Parts. But something else for the public? ebraker - do you change the disposition to "subsample" when the loan is closed? Maybe it would be better to change it to "used up"?
On Wed, May 18, 2022 at 5:06 PM Emily Braker @.***> wrote:
- [EXTERNAL]*
I really like the "subsampled" value because when an outside user sees "whole organism" in the part table and then "whole organism" a row below slightly indented, the "subsampled" status helps clarify what is going on. Also, many subsamples aren't in fact "in collection" but rather loaned/consumed, so regardless, the proposed change to update disposition to "in collection" may not be suitable.
— Reply to this email directly, view it on GitHub https://github.com/ArctosDB/arctos/issues/4688#issuecomment-1130694502, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADQ7JBDMGCGMBEXPHPF5JZ3VKVZY7ANCNFSM5WGIO24Q . You are receiving this because you were mentioned.Message ID: @.***>
"subsampled" status helps clarify what is going on
I don't think so - disposition should tell me where the thing is, that does not. WHAT the thing is - I think the intent - belongs - well, not there!
We can do WHATEVER with the UI, I'm always up for clever ideas.
Unrelated(ish), we should strive to banish the word "subsample." That's only one use case, any part can have a parent, there are lots of reasons to track derived parts beyond taking a small piece from a (probably piece of) liver.
Unrelated(ish), we should strive to banish the word "subsample." That's only one use case, any part can have a parent, there are lots of reasons to track derived parts beyond taking a small piece from a (probably piece of) liver.
Agree. Maybe we should be considering what is going to happen in DarwinCore where eventually I expect there to be a term "materialSampleParentID" in order to relate "child" parts with their "parent".
disposition should tell me where the thing is
HMMMM - maybe that should be part location? Container? Seems like we might have more than one way to say WHERE a part is?
OK, seriously. From Darwin Core - https://dwc.tdwg.org/terms/#dwc:disposition
The current state of a specimen with respect to the collection identified in collectionCode or collectionID.
Honestly not a great definition, especially when the examples include
in collection, missing, voucher elsewhere, duplicates elsewhere
duplicates elsewhere? how does that say anything about the state of the thing in question?
I guess what I am saying is what do we expect from "disposition"? OUR definition is as follows:
Disposition describes the status of parts and, as an abstract generality, the status of cataloged items.
This says nothing about WHERE and "subsampled" definitely seems like a "status". It seems like people are planning to get different information from this field, so maybe we are overloading it.
maybe that should be part location
Poor choice of words on my part - "can I get my hands on it?" or something might have been better - but yea, I've always thought that disposition would be a distraction and nothing else to anyone who REALLY uses machine-readable labels. If you know precisely where everything is and have good loan records and such you can probably derive the why (and whatever else disposition is supposed to do).
If humans are reading labels or typing, disposition seems complementary; I'd never fully trust eg part attribute location.
and, as an abstract generality, the status of cataloged items.
That's several years out of date; those data are still there, but they're all 'not applicable' and never exposed.
people are planning to get different information from this field,
Yes, that's the issue! I don't think I actually much care what's in disposition, but it is influencing how we see other things and that leads to impenetrable data. E.g., disposition is apparently seen as a "backup alternative" to transactions, which just PAINS me. (But, if the transaction really just isn't possible - which seems to happen - then maybe disposition is as good as anything that's avaialble.)
"subsampled" definitely seems like a "status"....people are planning to get different information
Maybe, but that's a structural characteristic of the same thing disposition is hooked onto; something designed to perform that job is RIGHT THERE. Maybe we need 'subsampled' to differentiate some other kind of parent-child relationship or something, but disposition (or my view of it) can't make sense for that either - things that are subsampled can still be lost, loaned, in collection, etc., and every part has exactly one disposition.
ANYWAY - agreed, we need better documentation, it seems on what disposition is, what it isn't, and maybe what it can approximate in desperate situations.
In the whole organism case, I'd advise indicating in the subsample part name what was actually taken - skin? a piece of liver? But for that, again, we need a way to select a different part name than the parent in the loan form.
On Wed, May 18, 2022 at 5:47 PM Mariel Campbell @.***> wrote:
I've been thinking for some time that it would be great if we had some means of automatically indicating when a part had been subsampled, when we check "subsample y/n?" on the loan forms. I subsample virtually everything, and I agree that the little indentation of the subsample is not obvious to anyone not doing this on a daily basis. But I haven't been changing the disposition to subsample - my subsamples are are consumable with disposition"on loan" . Maybe instead we could get a choice on the loan page to select the new part name (not currently an option - you get an indented "whole organism" or "heart, kidney, liver, spleen"). Some other way to indicate parent/child relationships? We have different part IDs, with indication of the parent part ID visible curatorially on Edit Parts. But something else for the public? ebraker - do you change the disposition to "subsample" when the loan is closed? Maybe it would be better to change it to "used up"?
On Wed, May 18, 2022 at 5:06 PM Emily Braker @.***> wrote:
- [EXTERNAL]*
I really like the "subsampled" value because when an outside user sees "whole organism" in the part table and then "whole organism" a row below slightly indented, the "subsampled" status helps clarify what is going on. Also, many subsamples aren't in fact "in collection" but rather loaned/consumed, so regardless, the proposed change to update disposition to "in collection" may not be suitable.
— Reply to this email directly, view it on GitHub https://github.com/ArctosDB/arctos/issues/4688#issuecomment-1130694502, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADQ7JBDMGCGMBEXPHPF5JZ3VKVZY7ANCNFSM5WGIO24Q . You are receiving this because you were mentioned.Message ID: @.***>
This needs a focused discussion with pre-approval by some group of people - I don't think that laying this down in the AWG will get us any resolution?
The funky data is almost certainly attracting more funky data, and we're coming up on two years of stalemate. I'm clearly unable to resolve this. If the AWG can't help, then who can? Is there just no resolution when someone or some group can't or won't find or accept a solution? That doesn't seem in any way sustainable.
I'm re-adding to the agenda in the hopes that the AWG can suggest SOMETHING, even if they can't or won't address the specifics.
ASNHC:Mamm records changed to transfer of custody and preparation method other with subsampled data in the method added.
@mvzhuang UTEP:Bird - changed to in collection, added a child part with disposition transfer of custody and preparation method other and subsample remark in method.
@mvzhuang UTEP:ES all were samples loaned out changed to used up (which was their condition) and changed condition to sample.
@jebrad UWBM:Mamm - changed to transfer of custody with subsample in condition - all were subsamples taken for loans.
From my work above, this probably doesn't make sense
update disposition="subsampled" to disposition="in collection"
My guess is that most of these either are a child part that should have a disposition of transfer of custody or used up OR they need a child part added. In the case of some of the above, I changed the child part type because it made sense to do that (UWBM child parts were hair not skin).
I'd be happy to work on the other things if given permission by the collections.
Thanks @Jegelewicz here's new data:
guid_prefix | count
-------------+-------
MSB:Mamm | 3
MVZ:Bird | 41
MSB:Para | 24
UAM:ES | 1
UNCG:Para | 4
DMNS:Bird | 1
MSB:Host | 17
DMNS:Mamm | 1375
MLZ:Bird | 3
UAM:Ento | 4
DMNS:Para | 14
UCM:Bird | 1
BYU:Herp | 5
UMNH:Mamm | 30
MSB:Herp | 10
UAM:Fish | 35
MSB:Bird | 3
MVZ:Mamm | 18
UCM:Herp | 6
UAM:Herb | 2
MMNH:Bird | 6
UNCG:Mamm | 20
UAMObs:Ento | 2
UCM:Mamm | 13
@ebraker @bryansmclean @mkoo @campmlc @ccicero @barke042 @camwebb/ @DerekSikes @cjconroy @jtgiermakowski @AdrienneRaniszewski @acdoll @jldunnum @jrdemboski @ccwlobo @byuherpetology @msbparasites @mlbowser @kderieg322079 @StefanieBond @catherpes @adhornsby
I would approve that anything "subsampled" gets a disposition of "used up". I can deal with the rare cases where a researcher actually sends us back any unused sample.
We actually do get unused subsamples returned fairly often. However, all our "subsamples" get recorded as disposition "on loan" until we hear back from the researcher that they are returning unused portions or that all samples were consumed. In that case, we change the disposition to either in collection or used up, and add part attributes and remarks as to remaining volume, condition report, and the fact the sample was returned from loan. It is part of our loan agreement that unused material is returned. It would be good to explicitly have "subsample" somewhere to indicate that status other than just using the indent which it not comprehensible to the general public. Condition seems like a good place for legacy info until we switch these values to a part attribute. Metadata would be useful.
Subsample also implies, but does not state explicitly, that the parent sample was handled in some way, potentially defrosted.
I agree that 'subsampled' doesn't seem like the best option for disposition. I'll need to look through these more closely, but I believe most of these tissue samples (liver, heart, etc) should be 'transfer of custody' or 'used up'. The 'whole organism' ones probably should have a different part name for these subsamples and then set to 'used up'. Then there are some without parent parts; I'll have to check on each of those individually. We'll get this cleaned up soon.
I think I got all of ours fixed. @dustymc can you run this query once more and see if I missed any?
guid_prefix | count
-------------+-------
MVZ:Mamm | 18
MMNH:Bird | 6
UCM:Mamm | 13
DMNS:Para | 12
UAM:Fish | 35
UAM:Herb | 2
DMNS:Mamm | 7
MSB:Mamm | 3
UAMObs:Ento | 2
UNCG:Para | 4
MSB:Herp | 10
UNCG:Mamm | 20
UAM:ES | 1
MSB:Host | 17
MLZ:Bird | 3
UCM:Bird | 1
UMNH:Mamm | 30
MSB:Bird | 3
UAM:Ento | 4
UCM:Herp | 6
MVZ:Bird | 41
MSB:Para | 24
(22 rows)
These get cleaned up, then used again almost immediately - perhaps this should be an issue to resolve at AWG Issues this week?
then used again almost immediately
That is a long-standing problem and seems to have happened here, we need a better approach.
guid_prefix | count
-------------+-------
MMNH:Bird | 6
UCM:Mamm | 13
UAM:Fish | 35
UAM:Herb | 2
MSB:Mamm | 3
UAMObs:Ento | 2
UNCG:Para | 4
MSB:Herp | 10
UNCG:Mamm | 20
UAM:ES | 1
MSB:Host | 1
MLZ:Bird | 3
UCM:Bird | 1
UMNH:Mamm | 30
MSB:Bird | 3
UAM:Ento | 4
UCM:Herp | 6
MVZ:Bird | 41
MSB:Para | 1