No abstract

Identifying novel drug-protein interactions is crucial for drug discovery. For this purpose, many machine learning-based methods have been developed based on drug descriptors and one-dimensional (1D) protein sequences. However, protein sequence can't accurately reflect the interactions in 3D space. On the other hand, a direct input of 3D structure is of low efficiency due to the sparse 3D matrix, and is also prevented by limited number of co-crystal structures available for training. In this work, we propose an end-to-end deep learning framework to predict the interactions by representing proteins with 2D distance map from monomer structures (Image), and drugs with molecular linear notation (String), following the Visual Question Answering mode. For an efficient training of the system, we introduced a dynamic attentive convolutional neural network to learn fixed-size representations from the variable-length distance maps and a self-attentional sequential model to automatically extract semantic features from the linear notations. Extensive experiments demonstrate that our model obtains competitive performance against state-ofthe-art baselines on the DUD-E, Human and Bind-ingDB benchmark datasets. Further attention visualization provides biological interpretation to depict highlighted regions of both protein and drug molecules.

A series of betulinic acid (BA) derivatives were designed and synthesized by introducing various fused heterocyclic rings at C-2 and C-3 positions. Their inhibitory effects of RANKL-induced osteoclastogenesis were evaluated by using a cell-based tartrate-resistant acid phosphatase (TRAP) activity assay. To our delight, most of these compounds exhibited a dramatic increase in inhibitory potency, compared with BA. The most potent compound, 20, showed 66.9% inhibition even at the low concentration of 0.1 μM, which was about 200-fold more potent than the lead compound BA. What’s more, the cytotoxicity assay on RAW264.7 suggested that the inhibition of 20 on osteoclast differentiation did not result from its cytotoxicity. The primary mechanistic study indicated that 20 could inhibit osteoclastogenesis-related marker gene expression levels of cathepsin K and TRAP. More importantly, 20 could attenuate bone loss of ovariectomy mouse in vivo. Therefore, these BA derivatives could be used as potential leads for the development of a new type of antiosteoporosis agent.

After new human transmissible H1N1 (swine flu) viruses were reported in Mexico and the United States in April 2009, the World Health Organization (WHO) announced the emergence of a novel influenza A virus. Most governments in the world have been alerted and are monitoring the situation closely. As one of the official responses to the H1N1 pandemic, the Chinese government has released three editions of a document entitled ''Recommended Schemes for Pandemic Influenza A Diagnoses and Treatments''. The third edition recommended the use of not only two targeted anti-flu drugs, oseltamivir and zanamivir, but also four anti-flu TCM (Traditional Chinese Medicine) prescriptions. Since then, TCM has played a significant role in fighting the pandemic. TCM drugs comprise multiple compounds regulating multiple targets for a given class of medical indications, and are tunable to the symptoms of the individual. This review summarizes anti-influenza agents from TCM, including compounds, herbs, and TCM prescriptions, and suggests that, by further investigating TCM theory and mining TCM databases, a better drug discovery paradigm may arise -one that can be beneficial to both TCM and modern medicine.

No abstract