chandkc / HTVS

0 stars 0 forks source link

Error in docking through LS algorithm #37

Closed LobeliaG closed 1 year ago

LobeliaG commented 1 year ago

machine- PIPL01 windows version- TEST_V2_5

PDBQT files were not found for the protein-ligand combinations when docking was performed using LS algorithm. for the same files, docking was successful using LGA. I used LS as algorithm, and reduced the no of runs to 1 to reduce the time. Other parameters were kept as it is. Input ligand was .csv file. image

chandkc commented 1 year ago

User cannot set the lower bound of LS RUN to 1 Same applies to LGA settings.

We will set the lower bound to 1 for both LGA and LS.

@sanvidat Please implement.

LobeliaG commented 1 year ago

When unbound model type is chosen as Extended, and the algorithm is LS, all of the combinations show error. For Bound type only some of them do. However for same inputs with LGA & extended model type, docking takes place alright & same as 'Bound' type unbound model. Number of runs were chosen more than 1 in each case. I am attaching the error files for both LS and LGA type produced from same input compound files. errorLGA.csv errorLS.csv input files- Input_HTVS.zip LSextended LS_para

chandkc commented 1 year ago

Extended state has some issue. It not functioning, as expected.

Work in progress . . . . . .

chandkc commented 1 year ago

We will mute the extended option, for now.

Needs time to underestand and then to fix.

LobeliaG commented 1 year ago

Extended option seems to be working okay with LGA. As both Bound & Extended model options with LGA & Bound model in LS are generating same results. Only the extended model in LS does not.

chandkc commented 1 year ago

Thank you @LobeliaG for this report.

I did check. There was something missing in the script of LS/Extended combination. Will rectify and release.

sed '/ga_pop_size/a set_ga' @sanvidat

sanvidat commented 1 year ago

fixed