The SAMPL9 logP challenge begins immediately and runs through Friday, Jan. 31 at 13:00 UTC. The compound list is provided here. Submissions must be uploaded via the submission server. A template file and an example file are provided in this directory, with additional instructions below.
For experimental measurements, we are indebted to Dr. Clara Ràfols (University of Barcelona), Dr. Rebeca Ruiz (Pion, UK) and William Zamora (University of Costa Rica), who measured the values used here and who are delaying publication of their experimental work to allow this challenge to proceed.
Experimental measurements were done for sixteen compounds with acid-base properties (see Table 1), most of them drugs showing a variety of therapeutical capabilities. These were commonly selected from Sigma-Aldrich ≥ 98% (www.sigmaaldrich.com), with some selected because of potential internal hydrogen bonding. The partition solvent was from Sigma-Aldrich: toluene (ACS reagent, ≥ 99.5%, 179418).
Experimental work is described in a manuscript published in PCCP.
Table 1. Experimental equipment used in the determination of logPtol/w
| name | Exp. equipment |
|---|---|
| Albendazole | Pion SiriusT3 |
| Alprenolol | Pion SiriusT3 |
| Amitriptyline | Sirius D-PAS & GLpKa |
| Bifonazole | Pion SiriusT3 |
| Chlorpheniramine maleate salt | Sirius D-PAS & GLpKa |
| Epinephrine | Pion SiriusT3 |
| Fluphenazine dihydrochloride | Sirius D-PAS & GLpKa |
| Glyburide | Pion SiriusT3 |
| Imipramine hydrochloride | Sirius D-PAS & GLpKa |
| Ketoprofen | Pion SiriusT3 |
| Nalidixic acid | Sirius D-PAS & GLpKa |
| Paracetamol | Pion SiriusT3 |
| Pindolol | Pion SiriusT3 |
| Quinine Sirius | D-PAS & GLpKa |
| Sulfamethazine | Sirius D-PAS & GLpKa |
| Trazodone hydrochloride | Pion SiriusT3 |
Partition values, log Ptol/w, were obtained by sample titrations as typically described for aqueous pKa determination but in presence of various amounts of the partitioning solvent. The log Psolv/w was calculated by the difference between the aqueous pKa and the apparent poKa (pKa measured in presence of a partition solvent) at several phase ratios depending on the expected partition value. All measurements were taken at 25 °C, under an inert gas atmosphere, and at least three titrations were made for each compound (Avdeef, 1993; Avdeef, Comer et al. 1993). Several titrations were carried out at the measurement limit conditions of the potentiometric technique. In these instances, the cautions for log Poct/w determination previously described were considered (Ràfols et al., 2012).
References:
- Avdeef, A., Comer J., 1993. pH-Metric log P.II. Refinement of partition coeficients and ionization constants of multiprotic substances, J. Pharm. Sci. 82, 183-190
- Avdeef, Alex; Comer, John E. A.; Thomson, Simon J., 1993. pH-Metric log P. 3. Glass electrode calibration in methanol-water, applied to pKa determination of water-insoluble substances, Anal. Chem., 65, 42-49
- Ràfols, C., Bosch, E., Ruiz, R., Box, K. J., Reis, M., Ventura, C., Santos, S., Araújo, M. E., Martins F., 2012. Acidity and Hydrophobicity of Several New Potential Antitubercular Drugs: Isoniazid and Benzimidazole Derivatives, J. Chem. Eng. Data, 57, 330-338
- Zamora, W. J.; Viayna, A.; Pinheiro, S.; Curutchet, C.; Bisbal, L.; Ruiz, R.; Ràfols, C.; Luwue, F. J. Prediction of toluene/water partition coefficients in the SAMPL9 blind challenge: Assessment of machine learning and IEF-PCM/MST continuum solvation models.
The SAMPL9 log P Challenge consists of predicting the water-toluene partition coefficients of 16 molecules. Our aim is to evaluate how well current models can capture the transfer free energy of small molecules between different solvent environments through blind predictions. Challenge participants are asked to predict the difference in free energy for the neutral form between water and toluene.
Submissions must be in the required format, linked below. The submission file must begin with the characters "logP_" and end with ".csv" and must contain a transfer free energy (in kcal/mol) for all compounds, as well as a predicted statistical uncertainty (same units) and a predicted model uncertainty (same units). Note that the statistical uncertainty may be small, e.g. if your model has very low statistical error, and should estimate how much your prediction would differ if you ran again, or used different initial coordinates, or changed other minor factors. Your predicted "model uncertainty" should be an estimate of the overall accuracy you would estimate your model has (regardless of statistical error), e.g. if you expect your prediction will be within 2 kcal/mol, use that value.
compounds.csv: A list of the compounds to be considered in the challengecompounds_image.png: An image of the compoundslogP_prediction_template.csv: Template file for predictionslogP_template_example.csv: Example completed template for submissionexperimental_data: Experimental data for logP challengeAnalysis: Includes submissions and eventually analysis; for now includes a very preliminary first pass at statistics.