Absorption, distribution, metabolism, and excretion (ADME) studies are critical for drug discovery. Conventionally, these tasks, together with other chemical property predictions, rely on domain-specific feature descriptors, or fingerprints. Following the recent success of neural networks, we developed Chemi-Net, a completely data-driven, domain knowledge-free, deep learning method for ADME property prediction. To compare the relative performance of Chemi-Net with Cubist, one of the popular machine learning programs used by Amgen, a large-scale ADME property prediction study was performed on-site at Amgen. For all 13 data sets, Chemi-Net resulted in higher R2 values compared with the Cubist benchmark. The median R2 increase rate over Cubist was 26.7%. We expect that the significantly increased accuracy of ADME prediction seen with Chemi-Net over Cubist will greatly accelerate drug discovery.

We propose a method based on neural networks to accurately predict hydration sites in proteins. In our approach, high-quality data of protein structures are used to parametrize our neural network model, which is a differentiable score function that can evaluate an arbitrary position in 3D structures on proteins and predict the nearest water molecule that is not present. The score function is further integrated into our water placement algorithm to generate explicit hydration sites. In experiments on the OppA protein dataset used in previous studies and our selection of protein structures, our method achieves the highest model quality in terms of F1 score, compared to several previous studies.

No abstract

No abstract

Syntactic parsing of web queries is important for query understanding. However, web queries usually do not observe the grammar of a written language, and no labeled syntactic trees for web queries are available. In this paper, we focus on a query's clicked sentence, i.e., a well-formed sentence that i) contains all the tokens of the query, and ii) appears in the query's top clicked web pages. We argue such sentences are semantically consistent with the query. We introduce algorithms to derive a query's syntactic structure from the dependency trees of its clicked sentences. This gives us a web query treebank without manual labeling. We then train a dependency parser on the treebank. Our model achieves much better UAS (0.86) and LAS (0.80) scores than state-of-the-art parsers on web queries.