Search for: All records

Synopsis Basic science research, also called “curiosity-driven research,” is fundamental work done with no immediate economic goals but rather a focus on discovery for discovery’s sake. However, basic science research is often needed to seed more applied, economically oriented, research. Both basic and applied research efforts are important aspects of the “bioeconomy,” defined here as the contributions to the overall economy from various biology-related fields spanning everything from museum-based natural history research to agricultural food and material production to healthcare. Here, we propose that more collaborative efforts across federal granting agencies in a venture-capitalist-like “PO-driven model” can help drive applied innovation from collaborations facilitated by program officers (POs). POs from NSF, DOE, DARPA, USDA, NASA, and other federal agencies should seek out parallel and complementary research ideas from grantees and provide funds to build teams of researchers who may otherwise be unaware of one another. Researchers working in different fields may also be unaware that the different organisms they are studying independently may have evolved similar traits (i.e., convergent evolution) that POs may recognize and who can then facilitate novel research avenues connecting those independent researchers (we provide examples of some projects inspired by convergent evolution here). In this top-down approach to research funding, the US bioeconomy will be pouring fuel on the fire of scientific productivity in this country. 

Inter-helix contact prediction is to identify residue contact across different helices in α-helical integral membrane proteins. Despite the progress made by various computational methods, contact prediction remains as a challenging task, and there is no method to our knowledge that directly tap into the contact map in an alignment free manner. We build 2D contact models from an independent dataset to capture the topological patterns in the neighborhood of a residue pair depending it is a contact or not, and apply the models to the state-of-art method's predictions to extract the features reflecting 2D inter-helix contact patterns. A secondary classifier is trained on such features. Realizing that the achievable improvement is intrinsically hinged on the quality of original predictions, we devise a mechanism to deal with the issue by introducing, 1) partial discretization of original prediction scores to more effectively leverage useful information 2) fuzzy score to assess the quality of the original prediction to help with selecting the residue pairs where improvement is more achievable. The cross-validation results show that the prediction from our method outperforms other methods including the state-of-the-art method (DeepHelicon) by a notable degree even without using the refinement selection scheme. By applying the refinement selection scheme, our method outperforms the state-of-the-art method significantly in these selected sequences. 

Abstract Effective cellular signaling relies on precise spatial localization and dynamic interactions among proteins in specific subcellular compartments or niches, such as cell-to-cell contact sites and junctions. In plants, endogenous and pathogenic proteins gained the ability to target plasmodesmata, membrane-lined cytoplasmic connections, through evolution to regulate or exploit cellular signaling across cell wall boundaries. For example, the receptor-like membrane protein PLASMODESMATA-LOCATED PROTEIN 5 (PDLP5), a potent regulator of plasmodesmal permeability, generates feed-forward or feed-back signals important for plant immunity and root development. However, the molecular features that determine the plasmodesmal association of PDLP5 or other proteins remain largely unknown, and no protein motifs have been identified as plasmodesmal targeting signals. Here, we developed an approach combining custom-built machine-learning algorithms and targeted mutagenesis to examine PDLP5 in Arabidopsis thaliana and Nicotiana benthamiana. We report that PDLP5 and its closely related proteins carry unconventional targeting signals consisting of short stretches of amino acids. PDLP5 contains 2 divergent, tandemly arranged signals, either of which is sufficient for localization and biological function in regulating viral movement through plasmodesmata. Notably, plasmodesmal targeting signals exhibit little sequence conservation but are located similarly proximal to the membrane. These features appear to be a common theme in plasmodesmal targeting. 

Abstract Background Hidden Markov models (HMM) are a powerful tool for analyzing biological sequences in a wide variety of applications, from profiling functional protein families to identifying functional domains. The standard method used for HMM training is either by maximum likelihood using counting when sequences are labelled or by expectation maximization, such as the Baum–Welch algorithm, when sequences are unlabelled. However, increasingly there are situations where sequences are just partially labelled. In this paper, we designed a new training method based on the Baum–Welch algorithm to train HMMs for situations in which only partial labeling is available for certain biological problems. Results Compared with a similar method previously reported that is designed for the purpose of active learning in text mining, our method achieves significant improvements in model training, as demonstrated by higher accuracy when the trained models are tested for decoding with both synthetic data and real data. Conclusions A novel training method is developed to improve the training of hidden Markov models by utilizing partial labelled data. The method will impact on detecting de novo motifs and signals in biological sequence data. In particular, the method will be deployed in active learning mode to the ongoing research in detecting plasmodesmata targeting signals and assess the performance with validations from wet-lab experiments. 

In this paper, we present a novel training method based on Baum-Welch algorithm for hidden Markov models (HMM), named as Comprehensive HMM (CompHMM), which changes the traditional approach of training HMM from positive examples only to be able to utilize both positive and negative examples in training HMMs. By comparison, our method outperformed the standard Baum-Welch method and another HMM discriminative training method significantly through both synthetic and real data in membership prediction task. 

Abstract Numerous cell surface receptors and receptor-like proteins (RLPs) undergo activation or deactivation via a transmembrane domain (TMD). A subset of plant RLPs distinctively localizes to the plasma membrane-lined pores called plasmodesmata. Those RLPs include theArabidopsis thalianaPlasmodesmata-located protein (PDLP) 5, which is well known for its vital function regulating plasmodesmal gating and molecular movement between cells. In this study, we report that the TMD, although not a determining factor for the plasmodesmal targeting, serves essential roles for the PDLP5 function. In addition to its role for membrane anchoring, the TMD mediates PDLP5 self-interaction and carries an evolutionarily conserved motif that is essential for PDLP5 to regulate cell-to-cell movement. Computational modeling-based analyses suggest that PDLP TMDs have high propensities to dimerize. We discuss how a specific mode(s) of TMD dimerization might serve as a common mechanism for PDLP5 and other PDLP members to regulate cell-to-cell movement.