@SuperScienceGrl There certainly is a degree of BS out there and I found it very calming to write this: https://t.co/OHvuIVzm21
@Dr_GHill My comment was more that visualization is useful precursor to, but not substitute for, data analysis http://t.co/OHvuIVzm21
@BranVanChemist You should see how creatively data can be analysed in science of #DrugDiscovery http://t.co/OHvuIVQWTz
@curiouswavefn @jchodera Try binning by X and then averaging Y. See http://t.co/RIxRkK9zQf or http://t.co/irY0h7cNrV for full instructions.
@curiouswavefn @jchodera Try binning by X and then averaging Y. See http://t.co/RIxRkK9zQf or http://t.co/irY0h7cNrV for full instructions.
@Moriarty2112 @janhjensen @drg1985 Citation counts of articles 'starring' in drug-likeness critique of interest? http://t.co/RIxRkK9zQf
@stuartcantrill @fxcoudert The editors of JCAMD were happy to publish these critiques: http://t.co/DjbFlh2aQZ http://t.co/RIxRkK9zQf
@Impactstory @menzenski Look at number of cites of articles critiqued in http://t.co/RIxRkK9zQf to see why I ignore citation metrics
+100 MT @pwk2013 @abhik1368 In the prediction we need to think in terms of effect size rather than stat significance http://t.co/zE8dPdMkah
@abhik1368 @rguha In the prediction we need to think in terms of effect size rather than statistical significance http://t.co/dcxWHorZCO
@hmkallewaard @RealTimeChem These may be helpful although probably too late for your meeting http://t.co/OHvuIVzTRz http://t.co/YmNOab0Qim
@eperlste @collabchem @Dereklowe Post publication rev that's actually in the literature http://t.co/dz2C8Eainb | http://t.co/dcxWHorZCO |
@eperlste @collabchem @Dereklowe Post publication rev that's actually in the literature http://t.co/dz2C8Eainb | http://t.co/dcxWHorZCO |
@Roche Who decides who is 'qualified'? Am I 'qualified'? Here are some of my credentials: http://t.co/OHvuIVQWTz
@geernst There is also the question of relevance: http://t.co/OHvuIVQWTz
@andyextance @NatureChemistry Ref 18 got coned in searchlights in: http://t.co/OHvuIVQWTz . See also slide 5 of: http://t.co/YmNOabhTkm
@DrugableNews @curiouswavefn You may find this article helpful (and possibly even instructive): http://t.co/OHvuIVQWTz
@petermurrayrust See if your machines can find examples of correlation inflation: http://t.co/OHvuIVQWTz ?
@ChemProfCramer Grades and physicochemical properties are not only things to become inflated with time http://t.co/OHvuIVQWTz
@0pensorcery @algae @openscience Here's a bit of post-publication review (in paywall journal) I'm afraid that I did: http://t.co/OHvuIVQWTz
@janhjensen If you are into dodgy stats you may find this relevant: http://t.co/OHvuIVQWTz
@rravi Just an idea (at this stage). Are you familiar with: http://t.co/OHvuIVQWTz
@rnl_s @openscience @F1000Research I've no problem with reviewing openly although prefer to get publication from it http://t.co/OHvuIVQWTz
@EconSciTech @openscience Maybe not: http://t.co/OHvuIVQWTz
@NatRevDrugDisc Correlation is not yes/no and you need to look at beyond statistical significance to effect size: http://t.co/OHvuIVQWTz
@damianpattinson @mattjhodgkinson @lteytelman @mrgunn @Jackstilgoe @openscience That's not always the case: http://t.co/OHvuIVQWTz
@carmendrahl @davidkroll @NeysaNevins @LSEImpactBlog Thanks for link. Can still can use regular scientific channels: http://t.co/OHvuIVQWTz
@rapodaca @ChemBark Post publication review sometimes makes it into the literature: http://t.co/OHvuIVQWTz
@ChemBark There is also some very peculiar data analysis in the published literature: http://t.co/OHvuIVQWTz
@egonwillighagen Much cheminformatic analysis suffers from transformation of the continuous to the categorical. http://t.co/OHvuIVQWTz
@eperlste @timtriche @openscience Postpublication review is not always a blessing: http://t.co/OHvuIVQWTz
@ChemProfCramer Drug-likeness is sometimes over-rated (see http://t.co/OHvuIVQWTz )
@EdwardTufte @ideapharma Also correlations can be inflated. http://t.co/OHvuIVQWTz
@ideapharma You may enjoy this data cookbook: http://t.co/OHvuIVQWTz
RT @rguha: nice commentary (http://t.co/CJTAHP7Z) on great paper http://t.co/UUE8EFvZ #drugdiscovery #statistics #data
nice commentary (http://t.co/CJTAHP7Z) on great paper http://t.co/UUE8EFvZ #drugdiscovery #statistics #data
@Dereklowe Yes. Interesting paper by Peter Kenny et al. http://t.co/eYEV7OFb