Antibodies are immune system proteins that protect the host by binding to specific antigens such as viruses and bacteria. The binding between antibodies and antigens are mainly determined by the complementarity-determining regions (CDR) on the antibodies. In this work, we develop a deep generative model that jointly models sequences and structures of CDRs based on diffusion processes and equivariant neural networks. Our method is the first deep learning-based method that can explicitly target specific antigen structures and generate antibodies at atomic resolution. The model is a ''Swiss Army Knife'' which is capable of sequence-structure co-design, sequence design for given backbone structures, and antibody optimization. For antibody optimization, we propose a special sampling scheme that first perturbs the given antibody and then denoises it. As the number of available antibody structures is relatively scarce, we curate a new dataset that contains antibody-like proteins as a complement to the original antibody dataset for training. We conduct extensive experiments to evaluate the quality of both sequences and structures of designed antibodies. We find that our model could yield highly competitive results in terms of binding affinity measured by biophysical energy functions and other protein design metrics.

BackgroundThe aim of this randomized controlled study was to determine whether octreotide (OCT) or scopolamine butylbromide (SB) was the more effective antisecretive drug controlling gastrointestinal (GI) symptoms due to malignant bowel obstruction (MBO) caused by advanced ovarian cancer.MethodsNinety-seven advanced ovarian cancer patients with inoperable MBO were randomized to OCT 0.3 mg/day (OCT group, n = 48) or SB 60 mg/day (SB group, n = 49) for 3 days through a continuous subcutaneous infusion. The following parameters were measured: episodes of vomiting, nausea, dry mouth, drowsiness, and continuous and colicky pain, using a Likert scale corresponding to a numerical value (none 0, slight 1, moderate 2, severe 3) recorded before starting the treatment (T0) and 24 h (T1), 48 h (T2), and 72 h after (T3) and the daily quantity of GI secretions through the Nasogastric tube (NGT) during the period of study. One patient in the SB group is not included in any assessments since she withdrew consent prior to receiving any treatment because of rapidly progressing cancer.ResultsOCT significantly reduced the amount of GI secretions at T1, T2, and T3 (P < 0.05) compared with SB. NGT secretions significantly reduced at T1, T2, and T3 compared with T0 (P < 0.05) in the OCT group, while in the SB group, only at T3, NGT secretions significantly reduced compared with T0. OCT treatment induced a significantly rapid reduction in the number of daily episodes of vomiting and intensity of nausea compared with SB treatment. No significant changes were observed in dry mouth, drowsiness, and colicky pain after either drug. Continuous pain values were significantly lower in the OCT group than in the SB group at T2 and T3 (P < 0.05).ConclusionsAt the doses used in this study, OCT was more effective than SB in controlling gastrointestinal symptoms of bowel obstruction. Further studies are necessary to understand the role of hydration more clearly in such a clinical situation.

There is still a lack of relevant studies on surgical site infection (SSI) after emergency abdominal surgery (EAS) in China. This study aims to understand the incidence of SSI after EAS in China and discuss its risk factors. All adult patients who underwent EAS in 47 hospitals in China from May 1 to 31, 2018, and from May 1 to June 7, 2019, were enrolled in this study. The basic information, perioperative data, and microbial culture results of infected incision were prospectively collected. The primary outcome measure was the incidence of SSI after EAS, and the secondary outcome variables were postoperative length of stay, ICU admission rate, ICU length of stay, 30-day postoperative mortality, and hospitalization cost. Univariate and multivariate logistic regression were used to analyze the risk factors. The results were expressed as the odds ratio and 95% confidence interval. A total of 953 patients [age 48.8 (SD: 17.9), male 51.9%] with EAS were included in this study: 71 patients (7.5%) developed SSI after surgery. The main pathogen of SSI was Escherichia coli (culture positive rate 29.6%). Patients with SSI had significantly longer overall hospital (p < 0.001) and ICU stays (p < 0.001), significantly higher ICU admissions (p < 0.001), and medical costs (p < 0.001) than patients without SSI. Multivariate logistic regression analysis showed that male (P = 0.010), high blood glucose level (P < 0.001), colorectal surgery (P < 0.001), intestinal obstruction (P = 0.045) and surgical duration (P = 0.007) were risk factors for SSI, whereas laparoscopic surgery (P < 0.001) was a protective factor. This study found a high incidence of SSI after EAS in China. The occurrence of SSI prolongs the patient's hospital stay and increases the medical burden. The study also revealed predictors of SSI after EAS and provides a basis for the development of norms for the prevention of surgical site infection after emergency abdominal surgery.

Deep generative models have achieved tremendous success in designing novel drug molecules in recent years. A new thread of works have shown the great potential in advancing the specificity and success rate of in silico drug design by considering the structure of protein pockets. This setting posts fundamental computational challenges in sampling new chemical compounds that could satisfy multiple geometrical constraints imposed by pockets. Previous sampling algorithms either sample in the graph space or only consider the 3D coordinates of atoms while ignoring other detailed chemical structures such as bond types and functional groups. To address the challenge, we develop Pocket2Mol, an E(3)-equivariant generative network composed of two modules: 1) a new graph neural network capturing both spatial and bonding relationships between atoms of the binding pockets and 2) a new efficient algorithm which samples new drug candidates conditioned on the pocket representations from a tractable distribution without relying on MCMC. Experimental results demonstrate that molecules sampled from Pocket2Mol achieve significantly better binding affinity and other drug properties such as druglikeness and synthetic accessibility.