|
||||||
| Postdoctoral associate | ||||||
| Modeling & Simulation, Biological Systems, Procter & Gamble | ||||||
|
Development of a knowledge management system for reproductive toxicity to support analogues identification in silico. This post doctoral associate will apply bioinformatics/cheminformatics concepts to develop novel methods for reproductive toxicity prediction with particular focus on analogue identification. At this moment, the only toxicological data that we use in risk assessment is limited to what is available in databases. These databases mainly focus on storing quantitative information, like LC50 values. However, many toxicological studies published in the literature have not been migrated to databases, for a variety of reasons: limited curator resources, obscure journals, nonstandard tests, etc. In addition, there is an increasing number of papers being published on the elucidation of mechanisms of toxicity that are not being stored in databases at all, so the information is available only as free text. Current literature searching methods are manual, and thus not resource effective. Net, we do not benefit from all available information when we conduct a safety assessment or try to find an analogue to fill in data gap. In order to make the process of selecting analogs for read–across more robust, there is a need to automate searches in free text to find a broader range of toxicological data and information to increase the number of potential analogs/ automatically extracting lines of evidence (the mechanistic information) from free text that would be potentially missed in manual work processes. Identification of pathways leading to organ specific in vivo effects will be one of the central goals of the project thus there will be a need to connect to pathway databases developed elsewhere. Approach: To develop the text mining tool, one will need to construct ontology describing the toxicological/mechanistic information to design Natural Language Processing templates and that are then will be used for subsequent chemical safety assessments. The templates will be coded in the existing cheminformatic data management system Ambit to allow for a combined text/data/structure queries. Required skill set: PhD in Biology/Toxicology with solid background in bioinformatics, cheminformatics is not required but an asset OR PhD in Bioinformatics with solid background in reproductive biology. Programming in Java, PERL is an asset as well. |
||||||
|