Using machine learning to improve ensemble docking for drug discovery.

Ensemble docking has provided an inexpensive method to account for receptor flexibility in molecular docking for virtual screening. Unfortunately, as there is no rigorous theory to connect the docking scores from multiple structures to measured activity, researchers have not yet come up with effective ways to use these scores to classify compounds into actives and inactives. This shortcoming has led to the decrease, rather than an increase in the performance of classifying compounds when more structures are added to the ensemble. Previously, we suggested machine learning, implemented in the form of a naïve Bayesian model could alleviate this problem. However, the naïve Bayesian model assumed that the probabilities of observing the docking scores to different structures to be independent. This approximation might prevent it from achieving even higher performance. In the work presented in this paper, we have relaxed this approximation when using several other machine learning methods-k nearest neighbor, logistic regression, support vector machine, and random forest-to improve ensemble docking. We found significant improvement.

Transfection of mammalian cells using linear polyethylenimine is a simple and effective means of producing recombinant adeno-associated virus vectors.

We have developed a simple protocol to transfect mammalian cells using linear polyethylenimine (PEI). Our linear PEI protocol is as effective as commercial reagents in the transfection of HeLa cells and XDC293 cells, a derivative of HEK293 cells, but at a fraction of the cost. Greater than 90% of XDC293 cells and 98% of HeLa cells transfected using our method were positive for EGFP expression as determined by flow cytometery. Our protocol should be useful for many different applications such as large-scale production of recombinant protein and viruses, which requires transient transfection of mammalian cells in large batches. We have used this protocol to produce recombinant adeno-associated virus (AAV) in XDC293 cells and in HeLa cells. This requires transient expression of three adenovirus gene-products (E2A, E4orf6, and VA RNAs) as well as the AAV replication (Rep78, Rep68, Rep52, and Rep40) and capsid (VP1, VP2, and VP3) proteins. Production of a recombinant AAV that expresses green fluorescent protein was assessed by quantitative PCR and by transduction of HeLa cells. Linear PEI is a better transfection reagent than calcium phosphate for the production of recombinant AAV in both HEK293 and HeLa cells. In addition, when both HeLa and XDC293 cells were by our method, HeLa cells in the absence of E1A generated three-fold more recombinant AAV than XDC293 cells, which constitutively express E1A.

Quantitation of encapsidated recombinant adeno-associated virus DNA in crude cell lysates and tissue culture medium by quantitative, real-time PCR.

Recombinant AAV vectors are produced by transient transfection of mammalian cells. The virus is usually purified from a combination of lysed cells and spent culture medium by HPLC. We have developed a quantitative, real-time PCR assay for quantifying encapsidated single-stranded viral DNA (i.e. DNA-containing virions) in cell lysates and the spent culture medium. This requires extensive DNaseI digestion to reduce the amount of AAV replicative DNA, as well as plasmid and cellular DNA, to negligible amounts. To demonstrate the utility of this assay, we produced recombinant AAV in HeLa cells and five different types of 293 cells. We used primers to the EGFP transgene to detect the production of a recombinant AAV. We assayed the cell lysates and media by both our quantitative PCR assay and a functional transduction assay. The quantitative PCR assay data correlated well with the transduction assay data. Because this assay only requires standard PCR primers and SYBR Green I dye to detect the amplification of the PCR template, it will readily adapt to any target DNA sequence within the recombinant AAV genome. The recombinant AAV vector does not need to express a reporter gene, such as EGFP or beta-galactosidase in order to assay the amount of virus produced.