Published November 29, 2019 | Version v1
Journal article Open

Melting point prediction using group contribution method. MP-Notebook presentation.

  • 1. University of Plovdiv

Description

The methods for quantitative structure-property relationship (QSPR) are recommended for regulatory usage together with the experimental protocols for particular endpoints. The melting point (MP) of a molecule is one of the physicochemical properties required from REACH (Registration, Evaluation, Authorisation and Restriction of Chemical Regulation). MP value (experimental or computed) plays important role in determining other properties such as vapor pressure, liquid viscosity, boiling point and etc. Computer prediction of the melting point is a great challenge. In this work, we present a group contribution model for predicting MP of organic compounds developed with the help of in-house developed open-source software Ambit-GCM. The modeling workflow includes: data manipulation and representation, descriptor and fingerprint calculation, model building, model validation, result output and storage. Typically such a workflow is accomplished by a number of different software tools and sometimes it hinders the reproducibility of published model results due to version differences, commercial software usage etc. One of the emerging software solutions for such issues are so called “Electronic Notebooks”. The notebook is a combination of text, data and code chunks, assembling tools from software packages which makes easy to trace the steps of a particular modeling workflow. We present the first prototype of MP-Notebook for building and validation of QSPR models for melting points of organic compounds.

Files

test_MP.csv

Files (40.6 MB)

Name Size Download all
md5:303beb7690fa16601ed7dfc3931a2cf5
3.0 MB Download
md5:1761f101bce4f10895091c8f5225ef2d
38.4 kB Download
md5:dfe38b4ba7e9872ba250b6aed5161368
66.6 kB Download
md5:f98dcbde2e6df59bf69e04ab56b0bd45
7.6 MB Preview Download
md5:b3975db545461230557aa9eb5ba21f29
30.0 MB Preview Download