Fast and effective molecular property prediction with transferability map.

Effective transfer learning for molecular property prediction has shown considerable strength in addressing insufficient labeled molecules. Many existing methods either disregard the quantitative relationship between source and target properties, risking negative transfer, or require intensive training on target tasks. To quantify transferability concerning task-relatedness, we propose Principal Gradient-based Measurement (PGM) for transferring molecular property prediction ability. First, we design an optimization-free scheme to calculate a principal gradient for approximating the direction of model optimization on a molecular property prediction dataset. We have analyzed the close connection between the principal gradient and model optimization through mathematical proof. PGM measures the transferability as the distance between the principal gradient obtained from the source dataset and that derived from the target dataset. Then, we perform PGM on various molecular property prediction datasets to build a quantitative transferability map for source dataset selection. Finally, we evaluate PGM on multiple combinations of transfer learning tasks across 12 benchmark molecular property prediction datasets and demonstrate that it can serve as fast and effective guidance to improve the performance of a target task. This work contributes to more efficient discovery of drugs, materials, and catalysts by offering a task-relatedness quantification prior to transfer learning and understanding the relationship between chemical properties.

Root-aligned SMILES: a tight representation for chemical reaction prediction.

Chemical reaction prediction, involving forward synthesis and retrosynthesis prediction, is a fundamental problem in organic synthesis. A popular computational paradigm formulates synthesis prediction as a sequence-to-sequence translation problem, where the typical SMILES is adopted for molecule representations. However, the general-purpose SMILES neglects the characteristics of chemical reactions, where the molecular graph topology is largely unaltered from reactants to products, resulting in the suboptimal performance of SMILES if straightforwardly applied. In this article, we propose the root-aligned SMILES (R-SMILES), which specifies a tightly aligned one-to-one mapping between the product and the reactant SMILES for more efficient synthesis prediction. Due to the strict one-to-one mapping and reduced edit distance, the computational model is largely relieved from learning the complex syntax and dedicated to learning the chemical knowledge for reactions. We compare the proposed R-SMILES with various state-of-the-art baselines and show that it significantly outperforms them all, demonstrating the superiority of the proposed method.

Identification of endogenous small peptides involved in rice immunity through transcriptomics- and proteomics-based screening.

Small signalling peptides, generated from larger protein precursors, are important components to orchestrate various plant processes such as development and immune responses. However, small signalling peptides involved in plant immunity remain largely unknown. Here, we developed a pipeline using transcriptomics- and proteomics-based screening to identify putative precursors of small signalling peptides: small secreted proteins (SSPs) in rice, induced by rice blast fungus Magnaporthe oryzae and its elicitor, chitin. We identified 236 SSPs including members of two known small signalling peptide families, namely rapid alkalinization factors and phytosulfokines, as well as many other protein families that are known to be involved in immunity, such as proteinase inhibitors and pathogenesis-related protein families. We also isolated 52 unannotated SSPs and among them, we found one gene which we named immune response peptide (IRP) that appeared to encode the precursor of a small signalling peptide regulating rice immunity. In rice suspension cells, the expression of IRP was induced by bacterial peptidoglycan and fungal chitin. Overexpression of IRP enhanced the expression of a defence gene, PAL1 and induced the activation of the MAPKs in rice suspension cells. Moreover, the IRP protein level increased in suspension cell medium after chitin treatment. Collectively, we established a simple and efficient pipeline to discover SSP candidates that probably play important roles in rice immunity and identified 52 unannotated SSPs that may be useful for further elucidation of rice immunity. Our method can be applied to identify SSPs that are involved not only in immunity but also in other plant functions.

The Vigna unguiculata Gene Expression Atlas (VuGEA) from de novo assembly and quantification of RNA-seq data provides insights into seed maturation mechanisms.

Legume research and cultivar development are important for sustainable food production, especially of high-protein seed. Thanks to the development of deep-sequencing technologies, crop species have been taken to the front line, even without completion of their genome sequences. Black-eyed pea (Vigna unguiculata) is a legume species widely grown in semi-arid regions, which has high potential to provide stable seed protein production in a broad range of environments, including drought conditions. The black-eyed pea reference genotype has been used to generate a gene expression atlas of the major plant tissues (i.e. leaf, root, stem, flower, pod and seed), with a developmental time series for pods and seeds. From these various organs, 27 cDNA libraries were generated and sequenced, resulting in more than one billion reads. Following filtering, these reads were de novo assembled into 36 529 transcript sequences that were annotated and quantified across the different tissues. A set of 24 866 unique transcript sequences, called Unigenes, was identified. All the information related to transcript identification, annotation and quantification were stored into a gene expression atlas webserver (http://vugea.noble.org), providing a user-friendly interface and necessary tools to analyse transcript expression in black-eyed pea organs and to compare data with other legume species. Using this gene expression atlas, we inferred details of molecular processes that are active during seed development, and identified key putative regulators of seed maturation. Additionally, we found evidence for conservation of regulatory mechanisms involving miRNA in plant tissues subjected to drought and seeds undergoing desiccation.