# Lipophilicity Prediction

*English translation by SOTA – AI Community of the Russian original. Organisers who would like this translation removed can email sota.ai.community@gmail.com.*

*Source: the 2023/24 task collection of the National Technology Olympiad (NTO), "Artificial Intelligence" profile, Stage 2, pp. 88–89: [original PDF](https://ntcontest.ru/upload/problembooks-2324/13_%D0%98%D1%81%D0%BA%D1%83%D1%81%D1%81%D1%82%D0%B2%D0%B5%D0%BD%D0%BD%D1%8B%D0%B9%20%D0%B8%D0%BD%D1%82%D0%B5%D0%BB%D0%BB%D0%B5%D0%BA%D1%82.pdf#page=88).*

## Second qualifying stage

The tasks of the second qualifying stage focus on property prediction and the design of small organic molecules. These tasks are highly relevant to modern chemistry and biology, as they play a key role in the development of new drugs, materials and technological processes.

## Individual task

### Task IV.1.1. (100 points)

*Topics: data analysis, QSAR methods, chemical databases.*

#### Problem statement

Participants are required to develop an algorithm capable of predicting the lipophilicity of small organic molecules. In essence, this parameter is the partition coefficient between a non-polar and a polar phase, that is, the ratio of the solubility of a compound in 1-octanol to its solubility in water. Lipophilicity is an important parameter, and its calculation is an integral part of the development of new medicinal drugs.

#### Input format

A file containing the string representation (`Simplified Molecular Input Line Entry System, SMILES`).

#### Output format

A file with the `CSV` extension containing the predicted lipophilicity values of the molecules (each value on a new line).

*Example of input data (file fragment)*

```
CCOc1ccccc1O
Cc1cccc(C)n1
CC(C)C#N
```

*Program output (file fragment)*

```
1.73
1.44
0.23
```

#### Assessment criteria

Solutions will be assessed with the standard metric for regression tasks, the root mean square error (RMSE).

Link to the description of the metric: <https://scikit-learn.org/stable/modules/model_evaluation.html#mean-squared-error>.

#### Solution

Participants have access to a baseline solution from the task developers that uses the `scikit-learn` and `RDKit` libraries.

The intended pipeline of the model includes:

- generating a representation of the molecules that is suitable for use as input data for predictive models;
- predicting the target parameter from the generated representation.

An example solution (with comments) is available at: <https://disk.yandex.ru/d/unyk1lr18_zxiw>.