Paperonium

Drug discovery · PLoS Comput Biol · 2020

Two compounds “completely satisfy” malaria-drug guidelines, the paper says. Its own table shows only one does.

The paper saysLabMol-149 passesunder 500 on bothLabMol-149,ordinary parasites2.9× the limitLabMol-149,drug-resistant parasites509, just overLabMol-152, ordinary parasites49, passesLabMol-152,drug-resistant parasites78, passes1,45015501505001,5005,000the limitpassesover
The paper saysLabMol-149 passesunder 500 on bothLabMol-149,ordinary parasites2.9× the limitLabMol-149,drug-resistantparasites509, just overLabMol-152,ordinary parasites49, passesLabMol-152,drug-resistantparasites78, passes1,45025500the limitpassesover

Dose that halves parasite growth, in nanomolar, against the 500-nanomolar limit

The paper says

A compound’s strength is the dose that halves parasite growth: smaller is stronger. The guidelines the paper cites ask for under 500 nanomolar (a unit of concentration) against ordinary malaria parasites and drug-resistant ones alike. Its summary and conclusion say LabMol-149 and LabMol-152 both pass.

Its own data show

Its own table gives LabMol-149 a dose of 1,450 nanomolar against ordinary parasites, 2.9 times the limit. (The table uses micromolar, a unit 1,000 times larger, and lists it as 1.450.) Against drug-resistant ones it is 509, just over. Only LabMol-152, at 49 and 78, is under 500 on both. One compound passes, not two.

Each figure averages three repeat tests; the pale bar around each dot shows how much they typically varied. The 509’s bar straddles the line, but even the low end of the 1,450 bar, 926, is nearly twice the limit. The summary says the paper’s computer method found two new families of drug leads meeting the guidelines; the table supports one.

“The summary” means the paper’s abstract: the short summary printed at its top, which is the part most readers see.

Read the paper →Full technical write-up →

Deep Learning-driven research for drug discovery: Tackling Malaria · PLoS Comput Biol (2020) · doi:10.1371/journal.pcbi.1007025