Identifying biosynthetic gene clusters from genomic data is challenging, with many in-silico tools suffering from a high rediscovery rate due to their dependence on rule-based algorithms. Next generation sequencing has provided an abundance of genomic information, and it has been hypothesized that there are many undiscovered biosynthetic gene clusters within this dataset. Here, we aim to develop a machine learning tool, ML-Miner, that infers patterns that describe a biosynthetic gene cluster in an unbiased manner and, as such, enables the identification of new biosynthetic gene clusters from genomic data. To solve this challenging problem, we define a simpler one to predict the class of a known BGC. Specifically, ML-Miner receives as input the concatenation of sequences that are known or believed to be part of a biosynthetic gene cluster. Its task is to identify which class it belongs, i.e. NPRS, PKS terpene and RiPPs. ML-Miner is a machine learning tool that uses Natural Language Processing, dimensionality reduction, and supervised learning to identify novel biosynthetic gene clusters. BioVec is a biological word embedding that we use to transform protein sequences from the highly curated MIBiG database of characterized biosynthetic gene clusters into their respective continuous distributed vector representations. Because the resulting protein vectors are of high dimensionality, a supervised Uniform Manifold and Approximation algorithm was employed to transform the high dimensional vectors into a robust lower-dimensional representation, as evaluated by Silhouette analysis, Hopkins’ statistic, and trustworthiness analysis. The density-Based Spatial Clustering of Applications and Noise algorithm showed that the clusters identified from the low dimensional datasets mapped to biosynthetic gene cluster types, defined with high accuracy in the MIBiG database. A random forest classifier was then trained and evaluated using the low dimensional vectors. It was shown to classify each biosynthetic gene cluster from the MIBiG database with excellent performance metrics. Finally, the model's ability to generalize was evaluated using biosynthetic gene clusters from the antiSMASH dataset, an uncurated database containing uncharacterized biosynthetic gene clusters. The performance metrics were high, with a balanced accuracy of ~85%. After a hyperparameter search, the balanced accuracy rose to ~90%. This suggests that ML-Miner is a robust machine learning pipeline that can be used to identify novel biosynthetic gene clusters. Future development of a confidence score for classification and a workflow for processing bacterial genomes into gene clusters will significantly improve the utility of this tool.

<p>  </p> <p>Actinobacteria are a large and diverse group of bacteria that are known to produce a wide range of secondary metabolites, many of which have important biological activities, including antibiotics, anti-cancer agents, and immunosuppressants. The biosynthesis of these compounds is often highly regulated, with many natural products being produced at very low levels in laboratory settings. Environmental factors, such as small molecule elicitors, can induce the production of secondary metabolites. These elicitors can be natural products, including antibiotics or hormones, or synthetic compounds. The use of small molecule elicitors to induce the production of secondary metabolites has several advantages. First, addition of elicitors to fermentation media can result in increased titers of known natural products. Second, elicitors can enable the discovery of novel natural products typically produced at undetectable levels. In recent years, there has been a growing interest in the use of small molecule elicitors to induce the production of secondary metabolites from actinobacteria, especially for the discovery of “silent” natural products. In this work, we sought to expand on the method of chemical induction by utilizing oxytetracycline at a sub-MIC concentration to induce secondary metabolite production in Streptomyces. We have shown that translation-inhibiting antibiotics, specifically oxytetracycline, have a profound effect on the production of coeliomycin P1, actinorhodin, and calcium-dependent antibiotics (CDAs) in S. coelicolor and S. lividans. The expression of actinorhodin in S. lividans under these conditions is unique, unlike its counterpart, S. coelicolor, which can produce actinorhodin under standard conditions. In addition to the increased production of known secondary metabolites, we have also demonstrated the induction of BGCs in several other strains of Streptomyces, which were observed via LC-MS. </p> <p>In addition to exploring antibiotics as elicitors we have explored the traditional approach of natural product discovery by taking an bioactivity guided approach. Several strain that we isolated from soil collect of Hawaii were screened for activity against several pathogenic strains primarily looking for which strain will inhibit the growth of a. baumannii, which is an intriguing target because the rate of resistance to common antibacterial medication is rising and it’s membrane composition is vastly different compared to other gram negative bacterium like E.coli. From this preliminary screening 1 strain (Streptomyces sp. CS62) out of the 8 that tested exhibited the desired biological activity. The supernatant of Streptomyces sp .CS62 was processed and screen by LC-MS to gain insight on the type of molecules that Streptomyces CS62 could produce. Upon our initial screening process none of the masses observed in the mass spec were matched to knowns. However, after 2D NMR analysis and genomic analysis it was unveiled that Streptomyces sp. CS62 produces factumycin a known antibacterial agent that targets A.baumannii .This unfourtunate turn of events illustrates the issues with natural product discovery and the need to improve natural product databases.</p> <p>In conjunction to discovering a novel producer of factumycin we are also investigating the production of antifungal compounds from Staphylococcus lugdunensis  a commensal strain that modulates the microbiome by producing lugdunin. The supernatant collected of Staphylococcus lugdunensis  is exclusively being test against Candida auris due to the immense health risk it possess to society because of its innate resistance to many antifungal drugs and its ability to rapidly gain resistance to other classes of antifungals.</p> <p>In addition to exploring the influence of antibiotics on secondary metabolite production and using bioactivity as a guide to discovering antibiotics. We are evaluating the soils collected from unique environments as potential sources for novel natural products. Specifically, we are evaluating the biosynthetic potential of bacteria from ore-forming environments, specifically fluorspar and topaz mines. Soils from ore-forming environments tend have low pH, high saline content, low water holding capacity, and poor nutrient availability. Therefore, ore-forming environments pose a hostile environment for life. To date, no one has explored the natural product potential, or the bacterial diversity, exhibited in these harsh environments. To assess the bacterial diversity, bacteria were isolated from various ore-forming environments using a procedure that is selective for actinobacteria. Following bacterial isolation, genomic DNA was isolated and 16s rRNA gene sequencing was performed to gauge the type of bacteria that were isolated. To stimulate secondary metabolite production, bacteria were then subjected to 7 different media conditions. The supernatant collected from these media conditions were tested against ESKAPE pathogens utilizing the CTSI broth microdilution assay. LC-MS MS analysis was performed for samples exhibiting biological activity. GNPS molecular networking was then utilized to determine potential molecules present in each sample.  Through this process we were able to identify one strain, which we named Streptomyces sp. S1A that exhibited a board range of biological activity (anticancer and antibacterial) and possess a wide array of biosynthetic gene clusters ranging complex macrolides (PKS and NRPS) to terpenes. </p> <p>In summary this multifaced approach to natural product discovery may lead to the discovery of novel antibiotics, enable us to increase production of known or unknown antibiotics through chemical induction, and the characterization of metabolites from Streptomyces sp. S1A will shed insight on the biochemical potential of organisms that inhabit ore-forming environments </p>

No description available.

No description available.

No description available.