Sammanfattning: Transcriptional memory is phenomenon that has gained interest due to its potential impact on human health. With our work, we wanted to contribute to the understanding of its regulation. To do so, we performed genome-wide knockout screens in both budding yeast and human cells. For that, we developed experimental platforms selecting an easy- to-assay readout that allowed us to stratify mutants based on phenotype and scrutinize for mutants enriched or depleted in the corresponding strata. Saccharomyces cerevisiae has been used for many years in genetics mainly because it is a simple unicellular eukaryote that is very easy to manipulate genetically. This made it a very useful model to study fundamental processes of gene expression in eukaryotes, such as transcriptional memory. We have discovered a new layer of complexity to the regulation of this phenomenon in yeast, based on mRNA stability. Despite the great utility of yeast as a model, the translation of the results to humans is challenging. When studying human cellular models, the biology is closer to the real situation in a human being while keeping it still relatively easy to work with. However, there was still a big challenge: precise and easy genetic manipulation. The discovery of the bacterial CRISPR systems and its application to genetic manipulation of high eukaryotes has paved the path to a whole new era in research. It can be used to knock in or knock out genes or intergenic regions of interest, edition, overexpression, epigenetic modifications, and many more. It has potential not only for fundamental research but also for synthetic biology, gene therapy, diagnostics, personalized medicine, etc. In this thesis, we have used a CRISPR/Cas-generated genome-wide knockout pool to interrogate factors involved in transcriptional memory in human cells. We have identified two putative factors involved in the regulation of transcriptional memory in humans that are interesting candidates for further research. To validate and further investigate candidates identified in pooled screens, knockout cell lines are frequently produced. A common method for that, is introducing two adjacent DNA double strand breaks (DSBs), using for example CRISPR/Cas with two gRNAs, that often results in the loss of the region in between. We have uncovered the occurrence of unexpected on-target aberrations while using a dual guide CRISPR/Cas system to produce deletion knockouts. We have shown that these events can go unseen and affect the phenotype of the cells. We proposed a workflow for comprehensive analysis of deletion clones. Finally, in 2020, scientists across the word were urged to help in the SARS‐CoV‐2 pandemic. As the virus spread quickly, it was crucial to combine the therapeutic efforts with the development of diagnostic tools. The gold standard detection method, RT-qPCR is very sensitive and specific, but is expensive and requires specialized equipment not available in all contexts. In addition, the reagents providers promptly started struggling to meet the global demand. The scientific community responded developing a plethora of alternatives. Some of those methods are based on loop-mediated isothermal amplification (LAMP) of the viral RNA, which can be detected either by fluorescence or change in pH among others. Compared to RT-qPCR, this technic is faster and cheaper, with a simple readout that does not require specific equipment. However, it still required the extraction of the viral RNA, and the supply of related reagents was rapidly affected. Here, we developed a LAMP-based method to detect SARS‐CoV‐2 without any previous manipulation of the sample, using in-house produced enzymes. Our method performs comparatively to the commercially available options in terms of sensitivity and specificity, and it is compatible with the most commonly used sample carriers. All the plasmids for the production of the enzymes used are publicly available.