We introduce a new representation and feature extraction method for biological sequences. Named bio-vectors (BioVec) to refer to biological sequences in general with protein-vectors (ProtVec) for proteins (amino-acid sequences) and gene-vectors (GeneVec) for gene sequences, this representation can be widely used in applications of deep learning in proteomics and genomics. In the present paper, we focus on protein-vectors that can be utilized in a wide array of bioinformatics investigations such as family classification, protein visualization, structure prediction, disordered protein identification, and protein-protein interaction prediction. In this method, we adopt artificial neural network approaches and represent a protein sequence with a single dense n-dimensional vector. To evaluate this method, we apply it in classification of 324,018 protein sequences obtained from Swiss-Prot belonging to 7,027 protein families, where an average family classification accuracy of 93%±0.06% is obtained, outperforming existing family classification methods. In addition, we use ProtVec representation to predict disordered proteins from structured proteins. Two databases of disordered sequences are used: the DisProt database as well as a database featuring the disordered regions of nucleoporins rich with phenylalanine-glycine repeats (FG-Nups). Using support vector machine classifiers, FG-Nup sequences are distinguished from structured protein sequences found in Protein Data Bank (PDB) with a 99.8% accuracy, and unstructured DisProt sequences are differentiated from structured DisProt sequences with 100.0% accuracy. These results indicate that by only providing sequence data for various proteins into this model, accurate information about protein structure can be determined. Importantly, this model needs to be trained only once and can then be applied to extract a comprehensive set of information regarding proteins of interest. Moreover, this representation can be considered as pre-training for various applications of deep learning in bioinformatics. The related data is available at Life Language Processing Website: http://llp.berkeley.edu and Harvard Dataverse: http://dx.doi.org/10.7910/DVN/JMFHTN.

We describe the design, fabrication, and performance of a bioreactor that enables both morphogenesis of 3D tissue structures under continuous perfusion and repeated in situ observation by light microscopy. Three-dimensional scaffolds were created by deep reactive ion etching of silicon wafers to create an array of channels (through-holes) with cell-adhesive walls. Scaffolds were combined with a cell-retaining filter and support in a reactor housing designed to deliver a continuous perfusate across the top of the array and through the 3D tissue mass in each channel. Reactor dimensions were constructed so that perfusate flow rates meet estimated values of cellular oxygen demands while providing fluid shear stress at or below a physiological range (<2 dyne cm(2)), as determined by comparison of numerical models of reactor fluid flow patterns to literature values of physiological shear stresses. We studied the behavior of primary rat hepatocytes seeded into the reactors and cultured for up to 2 weeks, and found that cells seeded into the channels rearranged extensively to form tissue like structures and remained viable throughout the culture period. We further observed that preaggregation of the cells into spheroidal structures prior to seeding improved the morphogenesis of tissue structure and maintenance of viability. We also demonstrate repeated in situ imaging of tissue structure and function using two-photon microscopy.

Echocardiography is essential to cardiology. However, the need for human interpretation has limited echocardiography’s full potential for precision medicine. Deep learning is an emerging tool for analyzing images but has not yet been widely applied to echocardiograms, partly due to their complex multi-view format. The essential first step toward comprehensive computer-assisted echocardiographic interpretation is determining whether computers can learn to recognize these views. We trained a convolutional neural network to simultaneously classify 15 standard views (12 video, 3 still), based on labeled still images and videos from 267 transthoracic echocardiograms that captured a range of real-world clinical variation. Our model classified among 12 video views with 97.8% overall test accuracy without overfitting. Even on single low-resolution images, accuracy among 15 views was 91.7% vs. 70.2–84.0% for board-certified echocardiographers. Data visualization experiments showed that the model recognizes similarities among related views and classifies using clinically relevant image features. Our results provide a foundation for artificial intelligence-assisted echocardiographic interpretation.

Vital organs maintain dense microvasculature to sustain the proper function of their cells. For tissue- engineered organs to function properly, artificial capillary networks must be developed. We have microfabricated capillary networks with a biodegradable and biocompatible elastomer, poly(glycerol sebacate) (PGS). We etched capillary patterns onto silicon wafers by standard micro-electromechanical systems (MEMS) techniques. The resultant silicon wafers served as micromolds for the devices. We bond the patterned PGS film with a flat film to create capillary networks that were perfused with a syringe pump at a physiological flow rate. The devices were endothelialized under flow conditions, and part of the lumens reached confluence within 14 days of culture. This approach may lead to tissue-engineered microvasculature that is critical in vital organs engineering.