Evaluation of log P, pKa, and log D predictions from the SAMPL7 blind challenge

Teresa Danielle Bergazin,Nicolas Tielker,Yingying Zhang,Junjun Mao,M. R. Gunner,Karol Francisco,Carlo Ballatore,Stefan M. Kast,David L. Mobley

DOI: https://doi.org/10.1007/s10822-021-00397-3

2021-06-24

Abstract:Abstract The Statistical Assessment of Modeling of Proteins and Ligands (SAMPL) challenges focuses the computational modeling community on areas in need of improvement for rational drug design. The SAMPL7 physical property challenge dealt with prediction of octanol-water partition coefficients and p K a for 22 compounds. The dataset was composed of a series of N-acylsulfonamides and related bioisosteres. 17 research groups participated in the log P challenge, submitting 33 blind submissions total. For the p K a challenge, 7 different groups participated, submitting 9 blind submissions in total. Overall, the accuracy of octanol-water log P predictions in the SAMPL7 challenge was lower than octanol-water log P predictions in SAMPL6, likely due to a more diverse dataset. Compared to the SAMPL6 p K a challenge, accuracy remains unchanged in SAMPL7. Interestingly, here, though macroscopic p K a values were often predicted with reasonable accuracy, there was dramatically more disagreement among participants as to which microscopic transitions produced these values (with methods often disagreeing even as to the sign of the free energy change associated with certain transitions), indicating far more work needs to be done on p K a prediction methods.

biochemistry & molecular biology,biophysics,computer science, interdisciplinary applications

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to evaluate the prediction accuracy of octanol - water partition coefficient (log P), acid dissociation constant (pKa) and the value of octanol - water partition coefficient at pH 7.4 (log D) in the SAMPL7 blind - test challenge. Specifically, this study focuses on the prediction of these physical properties of 22 compounds, which are mainly a series of N - acylsulfonamides and their related bioisosteres.

### Main Objectives:
1. **Evaluate Prediction Accuracy**: By comparing the prediction results submitted by different research groups with experimental data, evaluate the prediction accuracy of octanol - water partition coefficient (log P) and acid dissociation constant (pKa).
2. **Method Comparison**: Analyze the performance of different prediction methods, including quantum mechanics (QM), molecular mechanics (MM), database lookup (DL), linear free - energy relationship (LFER), quantitative structure - property relationship (QSPR) and machine learning (ML) methods.
3. **Microscopic and Macroscopic pKa Prediction**: Explore how to predict macroscopic pKa through the calculation of microscopic pKa, and the differences in the performance of different methods in microscopic conversion.

### Key Findings:
- **log P Prediction**: The accuracy of octanol - water log P prediction in the SAMPL7 challenge is lower than that in SAMPL6, which may be due to a more diverse data set.
- **pKa Prediction**: Compared with SAML6, the accuracy of pKa prediction in SAMPL7 remains unchanged. However, although the macroscopic pKa values can generally be predicted reasonably accurately, in terms of microscopic conversion, there are large differences among different methods, and even differences in the sign of the free - energy change in some conversions.
- **Method Improvement**: The study emphasizes the need for further improvement of pKa prediction methods, especially in the prediction of microscopic conversion.

### Formula Summary:
- **log P Definition**:
\[
\log P=\log_{10}K_{\text{ow}}=\log_{10}\left(\frac{[\text{non - ionic solute}]_{\text{octanol}}}{[\text{non - ionic solute}]_{\text{water}}}\right)
\]

- **pKa Definition**:
\[
\text{pKa}=-\log_{10}K_a
\]
where \(K_a\) is the acid dissociation equilibrium constant.

- **Relationship between Microscopic pKa and Macroscopic pKa**:
\[
\Delta G_0^{jk}=-\Delta m_{jk}C_{\text{units}}\text{pKa}_{jk}
\]
where \(\Delta m_{jk}\) is the charge change from state \(k\) to state \(j\), and \(C_{\text{units}} = RT\ln 10\).

- **pH - Dependent Free - Energy Change**:
\[
\Delta G_{jk}(\text{pH})=\Delta m_{jk}C_{\text{units}}(\text{pH}-\text{pKa}_{jk})
\]

- **Calculation of Macroscopic pKa**:
\[
x_j^q(\text{pH})=\frac{\exp\left(-\Delta G_{j\in q,k}(\text{pH})/RT\right)}{\sum_i\exp\left(-\Delta G_{ik}(\text{pH})/RT\right)}
\]
\[
x_j^{q(1)}(\text{pH})=x_j^{q(2)}(\text{pH})
\]

Through the detailed analysis of these formulas and methods, the paper aims to provide an in - depth understanding of the prediction of important physical properties in drug design and promote the further development of related fields.