upheno
upheno copied to clipboard
obophenotype
BioGRID ORCS database (of human cell line CRISPR screens) needs phenotype terms
The BioGRID database:
https://thebiogrid.org/
has recently started curating phenotypes resulting from CRISPR screens in human cell lines and have compiled the results of their curation into the ORCS database:
https://orcs.thebiogrid.org/
They are currently assigning phenotypes (traits really) to the various screen hits to indicate what was observed during the screen (what was being screened for). These are the terms currently used:
cell cycle progression (APO:0000253) protein transport (APO:0000129) protein/peptide accumulation (APO:0000149) regulation of signal transduction phenotype (CMPO:0000040) resistance to chemicals (APO:0000087) resistance to virus toxin resistance (APO:0000215) Tumorigenicity viability (APO:0000111)
Many terms are coming from APO (Ascomycete phenotype ontology) used by SGD and others in the yeast/fungi community and with which BioGRID is familiar as they have curated many genetic interactions in yeast. BioGRID would like to request more appropriate terms to apply to their annotations, but it isn't clear if these terms should be added to HPO, CMPO, or some other ontology as these are cell-line-specific phenotypes. I'm creating this ticket to discuss the matter. I've requested that BioGRID submit relative phenotype terms, e.g. "decreased viability", so as to clarify the needs.
Thank you for addressing this issue. In BioGRID's new CRISPR ORCS resource (orcs.thebiogrid.org), we are capturing phenotypes observed in various high-throughput CRISPR screens. In order to capture the phenotypes observed in hundreds of different cell lines, we've had to use a collection of phenotype terms from the APO, CMPO or else terms that we created.
However, as has been pointed out above, they're mostly traits. I think these observable traits can be coupled with qualifiers, such as those found in PATO (e.g. increased, decreased), but in order to help out with this discussion, I'll just list those observable-qualifier combinations that we have come across so far.
The following observables can be coupled with abnormal/normal or increased/decreased: cell cycle progression (APO:0000253) protein transport (APO:0000129) protein/peptide accumulation (APO:0000149) regulation of signal transduction phenotype (CMPO:0000040)
The following ones tend to be coupled with increased/decreased: resistance to chemicals (APO:0000087) resistance to virus toxin resistance (APO:0000215) tumorigenicity viability (APO:0000111)
The HPO is probably not the right fit as the phenotypes are observed in cell lines from different species. Please let us know whether uPheno can accommodate these phenotypes or whether we need to figure something else out.
@RoseOughtred, @chris-grove already put this on the agenda for our next phenotype editors call; I am absolutely in favour for acommodating this request. What is you timeline? Do you have a pending deadline to get the terms in somewhere?
We realized we have some additional phenotype terms that will be in our next release.
The following observables can be coupled with abnormal/normal or increased/decreased: cell cycle progression (APO:0000253) protein transport (APO:0000129) protein/peptide accumulation (APO:0000149) protein/peptide distribution (APO:0000209) regulation of signal transduction phenotype (CMPO:0000040) autophagy ( APO:0000074) apoptosis (APO:0000155)
The following ones tend to be coupled with increased/decreased: resistance to chemicals (APO:0000087) resistance to virus resistance to bacteria toxin resistance (APO:0000215) tumorigenicity viability (APO:0000111)
It would be great if this could be sorted out within two months so we can add the phenotype terms to BioGRID ORCS. We are aiming to submit our manuscript in April and it would be good if we could include the standardized cell line phenotype terms in our resource and publication. Will this work with your timeline? Many thanks!
On Mon, Feb 3, 2020 at 4:10 PM Nico Matentzoglu [email protected] wrote:
@RoseOughtred https://github.com/RoseOughtred, @chris-grove https://github.com/chris-grove already put this on the agenda for our next phenotype editors call; I am absolutely in favour for acommodating this request. What is you timeline? Do you have a pending deadline to get the terms in somewhere?
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/obophenotype/upheno/issues/586?email_source=notifications&email_token=AOORACWK5BYFDPOWRQV7BWLRBCB5PA5CNFSM4KPKWOC2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEKVNBGI#issuecomment-581619865, or unsubscribe https://github.com/notifications/unsubscribe-auth/AOORACW7TFEUZZGCZVQWIULRBCB5PANCNFSM4KPKWOCQ .
@RoseOughtred We discussed this today and I think the consensus was that species-specific phenotype ontologies should be used for each species experiment. Can you list out, explicitly, all the phenotypes that you would need for your curation? From your statements above I gather it could be the following list:
abnormal cell cycle progression abnormal protein transport abnormal protein/peptide accumulation abnormal protein/peptide distribution abnormal regulation of signal transduction phenotype abnormal autophagy abnormal apoptosis
increased cell cycle progression increased protein transport increased protein/peptide accumulation increased protein/peptide distribution increased regulation of signal transduction phenotype increased autophagy increased apoptosis
decreased cell cycle progression decreased protein transport decreased protein/peptide accumulation decreased protein/peptide distribution decreased regulation of signal transduction phenotype decreased autophagy decreased apoptosis
increased resistance to chemicals increased resistance to virus increased resistance to bacteria increased toxin resistance increased tumorigenicity increased viability
decreased resistance to chemicals decreased resistance to virus decreased resistance to bacteria decreased toxin resistance decreased tumorigenicity decreased viability
Can you prune that list to just the terms you would need? Also, can you indicate which terms you would need for which species? Then we can look for existing terms and request new terms as needed in the respective ontologies.
Thank you for looking into this and sorry to not be very clear. So far we have human, mouse and fly CRISPR screens in our collection. The only reason we used yeast APO terms is because they're available, even if they're the wrong species for our collection. So far we have not included any genome wide CRISPR screens in yeast as none have been published yet with full datasets. The genome wide CRISPR screens are being carried out in human, mouse or fly cell lines, not in the whole organisms. For this reason, we need all of the above phenotypes to be available for cell line branches of each species (human, mouse, fly, and others later on), if that makes sense. Also, at some point PATO qualifiers like abnormal, increased and decreased were separate terms from the actual phenotype terms. Are the qualifiers now being combined as part of the phenotype terms or am I misunderstanding the post above? Please let me know if anything isn't clear and thanks again for trying to figure out the best approach.
@RoseOughtred Yes, we would make sure that the terms selected for your use would be appropriate for application to cell lines. As far as I'm aware, most phenotype ontologies, with the exception of APO, currently adopt the meaning of phenotype that means a relative quality with respect to reference such that all phenotypes are some sort of abnormality of some trait. Some ontologies have a "normal" branch, which is something that needs to be sorted out, but even those have combined the terms (pre-composed phenotypes) for trait and quality into a single term. This also stems from a long-standing issue on the definition of what "phenotype" means: to some it means any (or perhaps all) measurable traits of an organism, whereas to others it strictly means an observed trait that differs from some reference. I believe most phenotype ontologies today take the latter meaning.
We discussed the list of phenotype terms above and think it would be best to keep all of them rather than removing some. We will likely need to use several of these terms and hope that we'll be able to request new ones as needed. For now, we'd appreciate it if these terms could be present in the human phenotype ontology (HPO), mammalian MPO, and the Drosophila DPO. We assume that they will all be part of the UPheno ontology as well. Please let us know what the estimated timeline might be on adding these terms. They would be most helpful for our new CRISPR screen database. Thanks!
I've created an MP ticket to address these terms in the MP https://github.com/obophenotype/mammalian-phenotype-ontology/issues/3164
@RoseOughtred To be clear, you want all those terms in all three ontologies?
Yes, that is correct. Thanks!
On Thu, Mar 5, 2020 at 4:12 PM Nico Matentzoglu [email protected] wrote:
@RoseOughtred https://github.com/RoseOughtred To be clear, you want all those terms in all three ontologies?
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/obophenotype/upheno/issues/586?email_source=notifications&email_token=AOORACS5ZSVGIUPV6D2Q6RLRGAIVLA5CNFSM4KPKWOC2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEN65SIQ#issuecomment-595450146, or unsubscribe https://github.com/notifications/unsubscribe-auth/AOORACTZQZNSKUJNHVCMXBTRGAIVLANCNFSM4KPKWOCQ .
@nicolevasilevsky Will you take care of this for the HP side of things? I think many of these should already be there.
@sbello would it be ok if we all worked on your spreadsheet side by side (columns for DPO and HP?). This would allow us to make sure all of these are created in an already reconciled fashion.
That would be great @matentzn That was part of the reason I made it a Google sheet that anyone can edit.
We've discussed some of these at MGI and we're good with using terms from the cellular phenotype and homeostasis/metabolism phenotype branches. I think we will need to make new terms for the infection related phenotypes as our terms for these currently live under immune system and refer to the organism response rather than the cellular response.
@nicolevasilevsky Will you take care of this for the HP side of things? I think many of these should already be there.
Yes, will do.
@sbello, thanks for letting me piggy back on your spreadsheet! I was just about to ask to do that, then I saw @matentzn already did and you said yes - perfect, thanks!
@RoseOughtred For Drosophila/flies, you can request terms at this GitHub repo:
https://github.com/FlyBase/drosophila-phenotype-ontology
Thanks @chris-grove
Sorry yeah forgot to say that.. DPO is considerably smaller with less resources, so it would be good if new term request could go to DPO one by one, and only if they are needed rather than pre-emptively. Correct me if I am wrong @Clare72 @dosumis
@nicolevasilevsky How else was I going to crib off your notes :) Seriously though coordinating the MP and HP terms only makes sense.
Thanks for coordinating all of this. I hope this is an easy phenotype addition. Could we please have the following terms in the MP and HP as well?
Increased phagocytosis Decreased phagocytosis Abnormal phagocytosis
Hi Rose, the HPO focuses on clinical observations/measurements. Phagocytosis is something that can be measured (e.g., https://www.ncbi.nlm.nih.gov/pubmed/28668892), so these terms are in scope, but a number of the other suggestions were not. Maybe we can find a time to skype about this.
Sounds good. Can you suggest a phenotype ontology that would be better for our CRISPR screen phenotypes which are observed in human cell lines? If the HPO is clinical and mainly pertains to the whole organism level rather than to cell lines, then what other ontology do you recommend instead?
I do not know if there is one. Possibly we could add a branch to HPO if there is no other resource. But it would be good to do this in a systematic way rather than ad hoc -- could you send me an email at (reversed org dot jax at robinson dot peter) thanks
@RoseOughtred DPO is a small ontology of ~200 terms and we do not currently have the resources to add all of the terms you request, especially as they are not fly-specific phenotypes.
If there is a Drosophila-specific term that you need - please use the DPO tracker and we can consider adding it to DPO.
Can someone remind me of the reasoning behind using species-specific terms rather than upheno terms + species qualifiers? None of these phenotypes look species-specific.
