search

OpenSourceMalaria / OSM_To_Do_List

Action Items in the Open Source Malaria Consortium
82 stars 13 forks source link

How to determine/prioritise target compounds #551

Open mcoster opened 6 years ago

mcoster commented 6 years ago

I debated whether to put this in Series 4, but I think it's a more generic question...

Has OSM ever used any voting or survey type tools to aid in choosing/prioritising synthetic targets? A kind of "wisdom of crowds" arrangement. As I plan for next year's OSM project students, I'm debating the best way to select synthetic targets to a) maximise the utility of any compounds synthesised, and b) minimise duplication of efforts.

mattodd commented 6 years ago

Ah yes. We've tried. @miike built a very cool standalone one from scratch that we never deployed properly (this was several years ago) but which he might still have. See #90

During Mark Gardner's webinar (#386) a vote happened, too, using part of the platform the webinar was on, using an on-screen display of several molecules. @MedChemProf won that one...

We have no good generic solution, though Twitter has a the ability to run a poll.

The original idea, way back, was to allow people to submit molecule ideas easily, and for people to upvote them and discuss them. We have a version of this here, of course (Github issues), but no easy way to display them as a table/graphic.

mcoster commented 6 years ago

Thanks for the info @mattodd - I thought this would be up OSM's alley! I guess complexity of the decision-making process and number of variables are issues, eg. a synthetic group needs to take into account what they can achieve in a timeframe, resource constraints, chemical availability, plus most of us have our favoured areas. I think I'll put out some ideas and a call for feedback, and depending on the result, I might see how a Twitter poll does.

miike commented 6 years ago

The super low tech solution is to use Github in an issue and leave one comment that serves as a description of the compound/timeframe etc and then use 👍 to vote up certain comments. It's not pretty and the barrier to entry is having a Github account.

Otherwise we may be best off with a free/low cost survey tool. Happy to trial some if there's demand for this going forward.

mattodd commented 6 years ago

One of the nice things about a page of suggested targets that can be voted on is that the structures persist in case someone later wants to have a crack at something, and we don't duplicate suggestions. The aim of the DCNYS page is exactly this, but there is obviously no interactivity there.

Naturally there may also be usefulness is shorter-term, more ad hoc polls, in case we're trying to decide between a handful of similar targets.

MFernflower commented 6 years ago

I'd like to see a web-based voting system where the molecule is displayed alongside calculations preformed by SwissADME - Anyone know any coders who could lend a hand? @spadavec perhaps?

mcoster commented 6 years ago

That would be cool and useful. A one line description of what hypothesis is being tested would be good too.

miike commented 6 years ago

Had another thought that involves relatively little effort.

Could we repurpose Google Forms and use that?

mattodd commented 6 years ago

Nice idea. I guess down the line there is the possibility that that could talk relatively easily to the current Master List of actual molecules in OSM.