QSPR PREDICTION OF VAPOR-PRESSURE FROM SOLELY THEORETICALLY-DERIVED DESCRIPTORS

To date, most reported quantitative structure-property relationship (Q SPR) methods to predict vapor pressure rely on, at least, some empiric al data, such as boiling points, critical pressures, and critical temp eratures. This limits their usefulness to available chemicals and incu rs the time and expense of experimentation. A model to predict vapor p ressure from only computationally derived molecular descriptors, allow ing study of hypothetical structures, is described here. Several multi linear regressions and artificial neural network analyses were tested with a range of descriptors (e.g., topological and quantum mechanical) derived solely from computations on molecular structure data. From a set of 479 compounds, a linear regression with an r(2) of 0.960 was ac hieved using polarizibility and polar functional group counts as descr iptors. This new computationally based model also proves to be more ac curate and works over a wider range of compound classes than most prev iously reported models.