Cecilia Pennica
Cecilia has been a manuscript editor for the BMC series journals since March 2019. Before joining BioMed Central, Cecilia completed a PhD in Mechanistic Biology and a Bachelor degree in Biochemistry from the University of York. During her PhD she investigated the molecular interaction underpinning the segregation of a multidrug resistance plasmid in bacteria. She then joined Imperial College of London for a short post-doc, working on DNA replication in budding yeasts. Cecilia is passionate about science communication and promoting open access.
Highlights of the BMC series: August 2019
• Deep-learning model to predict drug-drug interaction • Onset of lung cancer in young and older patients • Issues with primary health care facilities for rural pregnant women in Nigeria • New tool to help carers of people with dementia towards the end of life • Corticosteroid injection for heel pain: when is it effective?
BMC Bioinformatics – Novel deep learning model for more accurate prediction of drug-drug interaction effects
The most efficient therapy to many diseases is often the result of a combination of drugs. Administration of more than one molecule in cases of cancer, hypertension, asthma and AIDS can lead to an increase in drug efficiency and decrease in toxicity and resistance by the target cells. Nevertheless, different molecules can also interact in a detrimental way for the body. Hence the importance of being aware if multiple drugs are prescribed to patients. Different studies have been carried out, trying to foresee possible drug-drug interactions inside the body. In vivo testing is often limiting because it requires a long time to screen different compounds. In silico studies, instead, have proved to be more powerful.
Computational methods, usually, take into account known characteristics of molecules, such as structure, target and side effects. An article published this month by Lee et al. in BMC Bioinformatics shows the formulation of a new deep learning model to predict drug-drug interaction, that takes into consideration three concepts: target similarity profiles (TSP), Gene Ontology (GO) term similarity profiles (GO term gives information about the gene function, where it acts inside the cell and in which biological process it is involved) and Structural similarity profiles (SSP). Three similarity profiles are traced for any given drug pair. These are then joined together to provide an estimation of the level of interaction between the two molecules. This approach allowed the authors to identify, with the same or greater accuracy, drug-drug interactions that were characterized using different methodologies.
No hay comentarios:
Publicar un comentario