David Nasonov

Part 1: Developed an aggregate model, combining results from 3 unsupervised learning algorithms (Iso Forest, DBSCAN, Autoencoder) to flag athletes suspected in growth hormone abuse based on biomarkers Part 2: Built a Liquid Chromatography - Mass Spectrometry raw data pre-processing pipeline. Used the pre-processed data to train a Siamese Neural Network model to identify if two samples come from the same person or not. AUC = 92.3%, F1-score = 0.91

Heat stress during flowering poses a significant threat to oilseed rape (Brassica napus) production, necessitating a comprehensive understanding of the molecular mechanisms underlying plant responses to heat stress. Oilseed rape is a globally important crop with diverse applications, including as a source of vegetable oil, animal feed, and biofuel. However, its susceptibility to heat stress presents a challenge to sustainable cultivation. In this study, the genetic factors mediating heat stress response during flowering in oilseed rape were investigated. Here we show that heat shock factors (HSFs) and heat shock proteins (HSPs) play pivotal roles in the plant's ability to withstand heat stress. Complementing previous studies in other species, our findings reveal the specific upregulation of HSP20-like chaperones superfamily proteins, CLPBs, GolS1, BAG6, HOP3 and MBF1C inn response to heat stress, underscoring their potential as targets for genetic modification to enhance heat tolerance in oilseed rape. These results lay the groundwork for further study of heat stress response in Brassica and the potential development of genetically engineered varieties with improved heat tolerance. By elucidating the genetic and molecular basis of heat stress response in oilseed rape, this study has broader implications for sustainable crop production in the face of escalating environmental challenges associated with climate change.

- Raw data from the IMPC, in the form of 180,000 individual csv files, containing information about a single analysis, assessing effect of knocking out a gene on a given phenotype parameter was collated and cleaned - Parameters were grouped to reduce parameter space (275 -->19) - A scalable and comprehensive SQL database was created to efficiently store and query the data(Architecture seen on GitHub) - R Shiny app created to visualise a) the effect of the knockout every genes on a parameter, and the effects of knocking out one particular gene on every parameter ( Figure below and more on the GitHub page) - Project received 85% mark

- Processed the data obtained from single cell RNA-Seq of three different types of pluripotent stem cells in mice (raw counts) - Visualised the data, using volcano plots, PCA and heatmaps (seaborn, matplotlib) - Downstream analysis including Differential Gene Expression, GO-term analysis, pathway mapping (kegg/reactome) was performed in python and bash - Using the results from the analysis, and the existing literature, key set of genes for pluripotent differentiation and their relationships were determined - A Boolean GRN was constructed, where by manipulating the presence or absence of 5 extracellular signals, the gene activation state was modelled accurately