Notes on curating Chemistry papers

[CuzickA]

@martin2urban has a list of candidate chemistry papers for curation which he will add in a comment below.

Keep curation simple, focus on take-home message in title (Tier 1 curation #67). Additional annotations can be added if required at a later date.

Most annotations will be pathogen phenotypes. There may be a lot of overlap with other model organism databases eg https://www.yeastgenome.org/ For PHI-base we want to focus on pathogenic organisms showing a chemistry phenotype. (e.g. not Saccharomyces cerevisiae).

(Note: PHI-base will focus on PHI phenotype publications which may include single species phenotypes rather than publications containing ONLY pathogen or host phenotypes with the exception of chemistry publications. This should be documented in the video slides that Jon will be creating).

Links to other chemistry sessions #51 #53 #99.

Still to think about

1. priority papers contain lab-based gene modification, this could be in the ref strain (e.g. PMID: 22314539 The T788G mutation in the cyp51C gene confers voriconazole resistance in Aspergillus flavus causing aspergillosis. #51) or another strain (e.g. PMID:31501141 Trichophyton rubrum Azole Resistance Mediated by a New ABC Transporter, TruMDR3. #53 assumption here is that this would be compared to this strain WT phenotype, although these controls not curated for pathogen phenotypes). _YES, this is what we will curate
2. is it possible to curate papers only containing information on natural variants? _NO, we will not curate these at the moment
   Issues here are i) it may not be clear what to compare the natural variant seq/phenotype to? Really need a WT control strain to compare variant with. ii) how to make the genotype? add natural seq variation as if altered genotype or just use strain name (this is essentially curating a WT which we do not do for pathogen phenotypes)? This comes back to a past discussion on EV (engineered variant) and NV (natural variant) which was not implemented.
3. once these decisions have been made they can be added to the FAQ document Update the 'Phenotype tips' / 'FAQ' documentation canto-docs#10
   Current text in this document reads as follows....
   "Chemistry phenotypes
   Create your single species genotype (either pathogen or host genotype). Annotate your genotype with a PHIPO term indicating ‘resistance to chemical X’, ‘sensitive to chemical X’ or ‘normal growth on chemical X’. The experimental evidence will usually be ‘cell growth assay’ and the actual chemical X assayed can be added using the experimental conditions ‘+ chemical X’. Where ‘chemical X’ is the chemical under observation in the study."

Might need to add some info on point 2) above. DONE below

"Please note that we are NOT curating chemistry papers that only contain natural sequence variant data. We need the Chemistry publication to contain gene modification data, for example amino acid substitution data or overexpression genotypes."

[CuzickA]

Based on review of #53, #51 and #99 it looks as though we are NOT curating chemistry papers that only contain natural sequence variant data. We need the Chemistry publication to contain gene modification data.

[CuzickA]

We discussed the 'Criteria for curating chemistry papers' in the PHI-base group meeting yesterday (AC, KHK, JS and MU).

We agreed on the following

• Keep curation simple, focus on the take-home message in the title. Additional annotations can be added if required at a later date.

• Most annotations will be pathogen phenotypes. There may be a lot of overlap with other model organism databases eg https://www.yeastgenome.org/ For PHI-base we want to focus on pathogenic organisms showing a chemistry phenotype. (e.g. not Saccharomyces cerevisiae).

• (Note: PHI-base will focus on PHI phenotype publications which may include single species phenotypes rather than publications containing ONLY pathogen or host phenotypes with the exception of chemistry publications. This should be documented in the video slides that Jon will be creating).

• Chemistry study types we are currently curating

Priority papers contain lab-based engineered gene modification, this could be in the reference strain or another strain. We do not currently curate WT controls for pathogen phenotypes (like we do for metagenotypes). The assumption is that the modified genotype phenotypes are then mentally compared to the WT pathogen strain (reference or other strain listed within genotype).

UPDATE requested by KHK after looking at sessions #51 and #53
Moving forward we should curate the WT pathogen strain controls with the term 'normal growth on XXX'. This will make the interpretation of phenotypes (resistance, sensitive, normal growth...) for altered genotypes clearer in the PHI5 web display. (Note the comments and fig # number information are not displayed in PHI5). Curating the single species WT pathogen phenotypes will be an exception for chemistry curation. **FURTHER UPDATE KHK decided we no longer need to curate these WT controls 01_03_2023**

• Chemistry study types we are not currently curating

Papers only containing natural variant data.
(We cannot curate these natural variant papers now as there are some issues in doing this accurately. It would be nice to be able to curate these papers in the future. Curating the WT controls in the paper types above may be a step towards this)
The issues here are
i) It may not be clear what to compare the natural variant sequence/phenotype to? We really need a WT control strain to ‘mentally’ compare the variant with.
ii) How to create the genotype? Add the natural sequence variation (if known) as if it were an altered genotype or just use strain name (this is essentially curating a WT which we do not do for pathogen phenotypes)? This comes back to a past discussion on EV (engineered variant) and NV (natural variant) which was not implemented.
Interspecies complementation experiments. (added 27_01_2023 see #117)

[CuzickA]

KHK also suggested when selecting Chemistry papers for curation, we should consider the most recent publications first and then work backward through the older ones. This will mean that we are capturing recent data, publication authors may still be around to answer any questions and the data available in these publications may be of higher quality than some of the older publications. The exception to this would be that we would also like to curate the first publication describing a chemistry target.

[CuzickA]

We cannot curate interspecies complementation expts see #117.
Added to 'Criteria for curating chemistry papers' above.

[CuzickA]

Current info

Criteria for selecting and curating chemistry papers
Keep curation simple, focus on the take-home message in the title. Additional annotations can be added if required at a later date.
Most annotations will be pathogen phenotypes. There may be a lot of overlap with other model organism databases eg https://www.yeastgenome.org/ For PHI-base we want to focus on pathogenic organisms showing a chemistry phenotype. (e.g. not Saccharomyces cerevisiae).
Chemistry study types we are currently curating
Priority papers contain lab-based engineered gene modification, this could be in the reference strain or another strain. Moving forward we will also curate the WT strain controls. (This will be an exception for single species chemistry curation annotations). The assumption is that the modified genotype phenotypes are then mentally compared to the WT pathogen strain (reference or other strain listed within genotype).
Chemistry study types we are not currently curating
Experiments containing interspecies complementation data. It is not possible to curate two species within a pathogen genotype eg M. graminicola CYP51 (MgCYP51) variants in Saccharomyces cerevisiae (PMID: 20305029).
Papers only containing natural variant data. The issues here are
i) It may not be clear what to compare the natural variant sequence/phenotype to? We really need a WT control strain to compare the variant with.
ii) How to create the genotype? Add the natural sequence variation (if known) as if it were an altered genotype or just use strain name (this is essentially curating a WT which we do not do for pathogen phenotypes)? This comes back to a past discussion on EV (engineered variant) and NV (natural variant) which was not implemented.