Detecting concurrency bugs can be challenging due to the intricacies associated with their manifestation. These intricacies correspond to identifying the methods that need to be invoked concurrently, the inputs passed to these methods and the interleaving of the threads that cause the erroneous behavior. Neither fuzzing-based testing techniques nor over-approximate static analyses are well positioned to detect subtle concurrency defects while retaining high accuracy alongside satisfactory coverage. While dynamic analysis techniques have been proposed to overcome some of the challenges in detecting concurrency bugs, we observe that their success is critically dependent on the availability of effective multithreaded clients. Without a priori knowledge of the defects, manually constructing defect-revealing multithreaded clients is non-trivial. In this thesis, we design an approach to address the problem of automatically generate clients for detecting concurrency bugs in multithreaded libraries. The key insight underlying our design is that a subset of the properties observed when the defects manifest in a concur-rent execution can also be observed in a sequential execution. The input to our approach is a library implementation and a sequential testsuite, and the output is a set of multithreaded clients that can be used to reveal defects in the input library implementation. Dynamic defect detectors can execute the clients and analyze the resulting traces to report various kinds of defects including deadlocks, data races and atomicity violations. Furthermore, the clients can also be used by testing frameworks to report assertion violations. We propose two variants of our design – (a) path-agnostic client generation, and (b) path-aware client generation. The path-agnostic client generation process helps in detection of potential bugs present in the paths executed by the input sequential testsuite. It does not attempt to explore newer paths by satisfying path conditions either by modifying the input or by scheduling the threads appropriately. The generated clients are used to expose deadlocks, data races and atomicity violations. Our analysis analyzes the execution traces obtained from executing the input sequential clients and produces a concurrent client program that drives shared objects via library methods calls to states conducive for triggering deadlocks, data races or atomicity violations. For path-aware client generation, our approach explores newer paths that are not covered by the input sequential testsuite to generate clients. For this purpose, we design a directed, iterative and scalable engine that combines the strengths of static and dynamic analysis to help synthesize both multithreaded clients and schedules that violate complex correctness conditions expressed by the developer. Apart from the library implementation and the sequential testsuite as input, this engine also accepts a specification of correctness as input. Then, it iteratively refines each client from the input sequential testsuite to generate an ex-ecution that can break the input specification. Each step of the iterative process includes statically identifying sub-goals towards the goal of failing the specification, generating a plan toward meeting these goals, and merging of the paths traversed dynamically with the plan computed statically via constraint solving to generate a new client. The engine reports full reproduction scenarios, guaranteed to be true, for the bugs it finds. We have implemented prototypes that incorporate the aforementioned ideas and validated them by applying them on 29 well-tested concurrent classes from popular Java libraries, including the latest version of JDK. We are able to automatically generate clients that helped expose more than 300 concurrency bugs including deadlocks, data races, atomicity violations and assertion violations. We reported many previously unknown bugs to the developers of these libraries resulting in either fixes to the code or changes to the documentation pertaining to the thread-safe behavior of the relevant classes. On average, the time taken to analyze a class and generate clients for it is less than two minutes. We believe that the demonstrated effectiveness of our prototypes in helping expose deep bugs in popular Java libraries makes the design, proposed in this thesis, a vital cog in the future development and deployment of dynamic concurrency bug detectors.

Certains extraits naturels sont très rares et onéreux. La tentation est alors forte pour les producteurs de matières premières ou intermédiaires d’avoir recours à des adultérations. Le mélange avec des extraits issus d’origines botaniques apparentées, d’origines géographiques différentes, l’ajout d’un composé synthétique présent dans l’extrait naturel ou d’un autre extrait végétal, l’utilisation de produits phytosanitaires réglementés, la mise en place de procédés d’extractions non standardisés, en sont quelques exemples. Ces différences de qualité sont de plus en plus difficilement décelables. Cette thèse a pour objet de mettre au point une méthodologie rapide, efficace et non ciblée. Une approche métabolomique en UHPLC-HRMS est développée pour identifier ces défauts ou pratiques frauduleuses pour les absolues destinées au secteur arômes et parfums. Cette mise en évidence est réalisée par la détection de métabolites marqueurs. Les absolues présentent un défi particulier : entre 50 et 95 % de l’extrait peut être constitué de composés non volatils, rarement décrits dans la littérature. La recherche d’authenticité de ces extraits est alors plus complexe que dans le cas d’un extrait volatil tel qu’une huile essentielle, dont la composition peut être plus facilement déterminée par des techniques analytiques éprouvées et l’utilisation de bases de données. Deux plantes emblématiques de la parfumerie ont été étudiées : la violette (Viola odorata) et la rose (Rosa damascena et Rosa centifolia). Des marqueurs d’origine géographique ont été identifiés dans le cas de la violette, et des marqueurs d’origine botanique dans le cas de la rose. / Some natural extracts are very scarce and expensive. The temptation is therefore very high for producers or brokers to resort to adulterations. Mixing of extracts from related botanical origins, from different geographical origins, addition of synthetic compounds with natural occurrence in the extract, or addition of another vegetal extract, use of phytosanitary products, non-standardized extraction processes, are some examples. The quality differences are more and more difficult to detect. The objective of this PhD study was to develop a fast, efficient and non-targeted methodology. Metabolomics approach in UHPLC-HRMS was developped to identify defects or fraudulent practices in absolutes used in flavour and fragrances. These identifications are realized by the detection of chemical markers. Absolutes are a great challenge: between 50 and 95 % of the extracts consist of non-volatile compounds; moreover these products are seldom described in literature. The quest for validation for authenticity is then much more complex than in cases of volatile extracts such as essential oils, whose composition can be more easily determined by robust analytical instruments and numerous databases. Two symbolic plants used in perfumery were studied: viola (Viola odorata) and rose (Rosa damascena and Rosa centifolia). Markers of French origin were identified for viola, and markers of R. centifolia were identified for rose. Their characterizations were nevertheless the fundamental limit for this technique being at trace level in the extract. This work demonstrated the performance and limitation of the non-targeted metabolomics approach on absolutes, which are specialties of perfumery.

Les microorganismes représentent la forme de vie la plus diverse et abondante sur Terre et jouent un rôle fondamental dans tous les processus biologiques. Cependant, du fait de la grande diversité des communautés microbiennes, la caractérisation fine des environnements complexes reste difficile par les approches moléculaires actuelles de PCR et de métagénomique. En effet, ces approches ne conduisent qu’à une caractérisation partielle des communautés et ne permettent pas systématiquement d’associer la structure des communautés aux fonctions métaboliques réalisées. L’approche de capture de gènes par hybridation appliquée à des échantillons métagénomiques complexes a démontré son intérêt pour révéler toute la diversité connue mais aussi inconnue des biomarqueurs fonctionnels ciblés, ainsi que pour enrichir leurs régions flanquantes sur quelques centaines de permettant en évidence des associations de gènes. Ainsi, les travaux de thèse ont visé à développer une nouvelle méthode de capture de gènes par hybridation capable d’enrichir de façon ciblée de larges régions génomiques à partir d’échantillons complexes, permettant ainsi de faire le lien entre structure et fonction des communautés microbiennes. Ces développements ont nécessité la détermination de sondes de capture, l’utilisation d’une méthode d’extraction d’ADN de haut poids moléculaire et la mise au point d’un protocole de capture permettant de piéger des fragments nucléiques de grande taille (jusqu’à 50 kb). La validation de la méthode de capture par hybridation sur un échantillon environnemental de sol a permis de révéler tout son potentiel. Appliquée au gène exprimant l’ARNr 16S, cette stratégie a permis de révéler une diversité microbienne non accessible par les approches moléculaires conventionnelles, avec une résolution d’identification jusqu'au niveau de l’espèce rendue possible grâce à la reconstruction de la séquence complète de ce marqueur phylogénétique. Appliquée à un gène fonctionnel, elle a conduit à la reconstruction de la séquence du biomarqueur et de ses régions flanquantes pouvant atteindre plusieurs dizaines de kb, permettant d’identifier les microorganismes possédant les capacités métaboliques d’intérêt. Ainsi, la capture par hybridation représente une approche alternative prometteuse pour le diagnostic environnemental en conduisant à une meilleure caractérisation des communautés microbiennes. / Microorganisms are the most diverse and abundant life forms on Earth and are key players in thefunctioning of all biological processes. Nevertheless, PCR and metagenomics strategies aiming to describemicrobial communities are hampered by their huge diversity. Indeed, these molecular methods only drive to apartial description of communities and do not systematically allow linking functions back to the identities of themicroorganisms. Hybridization capture applied to complex metagenomic samples has demonstrated its efficiency to reveal all known and unknown diversity of targeted biomarkers, and to enrich their flanking regions over a few hundred bp facilitating the discovery of gene associations.Thus, this work aimed at developing a new hybridization capture method capable of specifically enrichinglarge genomic regions from complex samples allowing to associate structure and functions of communities. Thedevelopment of this method required the design of capture probes, the use of a high molecular weight DNAextraction method, and the elaboration of a capture protocol dedicated to the enrichment of large genomicfragments (up to 50 kbp).The validation of the hybridization capture method on an environmental soil sample uncovered all itspotential. Applied to the 16S rRNA gene, this strategy revealed greater microbial diversity than conventionalmolecular methods and improved phylogenetic resolution up to the species level thanks to the reconstruction offull-length genes. Applied to a functional gene, the method enabled the reconstruction of large genomic regionscarrying the targeted biomarker and its flanking regions over several tens of kbp, leading to the identification ofmicroorganisms with specific metabolic functions. Hybridization capture thus appears as a promising alternativemethod for environmental diagnosis, through providing a better knowledge of microbial communities.

The rationale behind targeted molecular therapy in cancer, oncogene addiction, is that tumors rely on driver oncogenes to control their proliferation and survival. Therefore, an efficacious targeted therapy should induce a dual, detrimental response to the tumor. While there have been clinical success stories using targeted therapies, even tumors that are initially sensitive invariably develop resistance. In the case of triple negative breast cancer (TNBC), despite extensive evidence pointing to its driver oncogene status, inhibitors of the Epidermal Growth Factor Receptor (EGFR) are considered clinically inefficacious. Resistance to EGFR inhibition has been predominantly described as due to genetic alterations. Yet it remains unclear why patients exhibiting the same dysregulated status of a driver oncogene react to targeted therapy, as in the case of EGFR-mutant non-small cell lung cancer, while others do not at all (i.e., TNBC). Furthermore, not all of resistance can be described by genetic alterations to EGFR, to its pathway effectors, or to compensatory pathways. Emerging data reveals that drugs can induce resistance by rewiring epigenomic, transcriptional, and translational regulatory mechanisms. Unfortunately, a major limitation in designing efficacious treatments is our inability to predict whether cell types can rewire in response to drug exposure. Therefore, it is necessary to elucidate mechanisms of growth and survival in cells that have undergone rewiring. This study characterized intrinsic resistance to EGFR inhibition in TNBC. We found that EGFR inhibition induces rewiring, which results in a resistant growth state that bypasses the EGFR-MAPK pathway as a whole. Additionally, we found that a tRNA-modifying complex masks the oncogene addiction status of EGFR in TNBC by stabilizing the protein abundance of a pro-survival protein. Importantly, this happens solely in the context of EGFR inhibition. Taken together, this study highlights potential therapeutic strategies for TNBC and strategies that can be used to improve our understanding of targeted therapy resistance, especially intrinsic resistance.

In 2013 the California Legislation passed a new K-12 School accountability mandate. The Local Control Accountability Plan was sought to increase the educational equity for targeted student groups in addition to allowing school districts to mine a diverse set of local school data to develop goals in the 8 priority areas that speak to the needs of their local students. A requirement of the LCAP was that school districts include a diverse set of stakeholders to work in a collaborative manner to develop, critique, and refine local goals. Stakeholder groups are required to consist of district-level administrators, teachers, staff, students, parents, and community stakeholders. This qualitative study focused on the experiences and insights of district and community stakeholders as they participated in the development of LCAPs. To answer the questions and understand the experience of stakeholders interviews, document analysis, and observations were used. The results of the study revealed that while the intent of bringing a diverse set of stakeholders is noble there is a need to build trust amongst stakeholders. Findings also show the need to build community stakeholder understanding of the purpose, promises, and limits of the LCAP. The study revealed the need for school districts to examine and dismantle historic practices in order to fully embrace and implement the LCAP mandates.

Tato práce je zaměřena jak na cílený tak na přehledný přístup ve studiu proteomiky. Cílená proteomika přináší informace o přítomnosti proteinu a jeho lokalizaci v buňce či tkáni pomocí luminiscenčních značek na bázi kvantových teček, zatímco přehledná proteomika se zabývá identifikací změn v proteomu dvou nebo více jedinců stejného druhu vystavených různým podmínkám. Protože proteomika vyžaduje vysoce citlivé separační a identifikační techniky, byly v této práci ověřeny různé metody zlepšení citlivosti kapilární elektroforézy s hmotnostní detekcí. Použití rozhraní s kapalinovým spojem pro spojení těchto dvou technik, které zajišťuje vyšší citlivost analýz, bylo také ověřeno analýzou metabolitů etanolu a kokainu v lidské moči. Zavedené techniky instrumentace jsou využitelné při posouzení vlivu významných faktorů životního prostředí na živé systémy jak na buněčné tak na molekulární úrovni.

Les tumeurs sont généralement constituées de différentes sous-populations de cellules cancéreuses génétiquement hétérogènes, responsables en grande partie de la capacité de la tumeur à évoluer rapidement et à s’adapter aux conditions environnementales. Cette diversité génétique a des conséquences majeures pour le patient, notamment au cours de la progression tumorale et pour l’acquisition d’une résistance aux traitements.Nous avons développé une nouvelle stratégie basée sur la technologie CRISPR/Cas9 qui consiste à introduire, en plus d’une altération de séquence voulue d’un gène d’intérêt, une série de mutations silencieuses, constituant une sorte d’étiquette génétique qui peut être détectée par PCR quantitative ou séquençage de nouvelle génération. En parallèle, un code-barres constitué exclusivement de mutations silencieuses est utilisé comme contrôle interne pour les effets non spécifiques potentiels qui peuvent être engendrés suite au clivage hors-cible par le système CRISPR/Cas9. Cette approche, que nous avons appelée CRISPR-barcoding, permet de générer et de suivre l’émergence d’un petit groupe de cellules cancéreuses contenant une mutation voulue au sein d’une population de cellules non modifiées, représentant ainsi un nouveau modèle expérimental d’hétérogénéité génétique intratumorale. Grâce à une série de preuves de concept, nous avons montré que CRISPR-barcoding est une nouvelle approche qui permet d’étudier de façon simple et rapide les conséquences fonctionnelles de différents types de modifications génétiques apportées directement au niveau de la séquence génomique.Dans la deuxième partie de ma thèse, nous avons utilisé cette nouvelle approche pour l'étude de la résistance du cancer bronchique non à petites cellules (CBNPC) à la thérapie ciblée. Les patients de CBNPC dont la tumeur présente une mutation activatrice de l'epidermal growth factor receptor (EGFR) sont généralement traités avec des inhibiteurs de ce récepteur. Malheureusement, malgré une réponse initiale, presque invariablement ces patients rechutent, suite au développement d’une résistance causée, dans la majorité des cas, par l’apparition de mutations secondaires ou tertiaire de l'EGFR. Grâce à un criblage réalisé en utilisant un modèle de CBNPC basé sur la stratégie CRISPR-barcoding, nous avons pu identifier l’inhibiteur multikinase sorafenib pour sa capacité à prévenir la résistance de ces cellules aux inhibiteurs d’EGFR. Ce composé présente un mécanisme d’action original, impliquant une diminution précoce du niveau de phosphorylation de STAT3, suivie par une baisse considérable de l’expression de l’EGFR, aboutissant à une inhibition des voies intracellulaires en aval de ce récepteur, telles que RAS-MAPK et PI3K-AKT-mTOR. Ces données ont été confirmées in vivo en utilisant un modèle de xénogreffe de cellules de CBNPC modifiées par CRISPR-barcoding.En conclusion, l’ensemble de nos résultats montre que le sorafenib peut prévenir l’émergence de cellules de CBNPC résistantes aux inhibiteurs d’EGFR, indiquant que ce composé pourrait représenter une nouvelle stratégie thérapeutique pour le traitement de ce type de tumeur. / Individual tumors are composed of multiple and genetically distinct subpopulations of transformed cells that can adapt and evolve in a different way based on environmental conditions. This genetic diversity has major consequences for the patient, particularly during tumor progression and for cancer treatment.We devised a new strategy based on CRISPR/Cas9 technology in which a potentially functional modification in the sequence of a gene of interest is coupled with a series of silent point mutations, functioning as a genetic label for cell tracing. In parallel, a second barcode consisting of distinct silent mutations is inserted in the same cell population and used as a control for CRISPR/Cas9 off-target cleavage. This approach, that we named CRISPR barcoding, enables detection of cells containing the mutation of interest within a mass population of unmodified cells using real-time quantitative PCR or deep sequencing. Through a series of proof-of-concept studies, we demonstrated that CRISPR-barcoding is a fast and highly flexible strategy to investigate the functional consequences of a specific genetic modification in a broad range of assays.In the second part of my thesis, we used CRISPR-barcoding to investigate non-small cell lung cancer (NSCLC) resistance to targeted therapy. Some NSCLCs harbor activating mutations of the epidermal growth factor receptor (EGFR) and are addicted to this signaling pathway. These tumors initially show a good response to EGFR inhibitors (EGFRi), but they almost invariably relapse, due to the acquisition of a resistance, as a result of additional genetic alterations, including secondary and tertiary EGFR mutations. Using a CRISPR-barcoding model, we identified the multikinase inhibitor sorafenib for its ability to prevent EGFRi resistance in NSCLC cells. This compound acts through an original mechanism that involves early reduction of STAT3 phosphorylation and late down-regulation of EGFR, resulting in the inhibition of different downstream pathways activated by this receptor, including, RAS-MAPK and PI3K-AKT-mTOR. These results were confirmed in vivo, using a CRISPR-barcoding xenograft model for NSCLC.Altogether, our data indicate that sorafenib can prevent NSCLC resistance to EGFRi through a novel mechanism, thus providing a new potential therapeutic strategy for the treatment of this type of cancer.

Le biofilm bactérien constitue un obstacle majeur à la cicatrisation des plaies. Par ailleurs, il est responsable de l’émergence d’une résistance et d’une tolérance accrues aux antibiotiques. Par conséquent, le développement de systèmes de délivrance contrôlée d’un agent ciblant la structure du biofilm apparaît comme une approche thérapeutique alternative indispensable et urgente pour la prise en charge des plaies chroniques. A travers cette étude, nous avons développé des systèmes membranaires pour pansements libérant une protéine, la dispersine B (DB),capable de cibler de manière sélective la matrice du biofilm, en créant un microenvironnement délétère pour le biofilm bactérien. Pour ce faire, nous nous sommes intéressés aux membranes asymétriques (MAs) à base de polyesters biodégradables tels que le poly(3-hydroxybutyrate-co-4-hydroxybutyrate), le poly (butylène succinate-co-butylène adipate) (PBSA), et l’acide polylactique. En incorporant dans la solution de polymère des agents porogènes hydrophiles (APs), nous avons pu obtenir des MAs à porosité élevée, un réseau poreux interconnecté, perméables au dioxygène et à l’eau vapeur. En utilisant l’albumine de sérum bovin, nous avons pu montrer que la capacité de piégeage de la protéine et sa libération contrôlée à partir des MAs de PBSA était influencée par la structure de celles-ci et la présence d’APs résiduels. Les études in vitro ont montré une très grande efficacité anti-biofilm à la fois en inhibition et en dispersion (jusqu’à 80%). Les tests standards normalisés de cytotoxicité in vitro ont montré que les MAs de PBSA non chargées et chargées en DB répondaient aux critères de cytocompatibilité exigées pour une application de type pansement. / Bacterial biofilms are a major obstacle to the wound healing process. In addition, they are responsible for the emergence of resistance and tolerance to antibiotics. Hence, the development of controlled drug delivery systems targeting the bacterial biofilm appears as an urgent and essential alternative therapeutic approach for the effective management of chronic wound. In this work, we developed wound dressings in which a protein, dispersin B (DB), is released capable of selectively targeting the biofilm matrix, creating a deleterious microenvironment for the bacterial biofilm. To this end, we were interested in asymmetric membranes (AMs) from biodegradable polyesters such as the poly(3-hydroxybutyrate-co-4-hydroxybutyrate), the poly (butylene succinate-co-butylene adipate) (PBSA) and the polylactic acid. By the incorporation of hydrophilic porogen agents (PA), we were able to obtain AMs with a high level of porosity, exhibiting a porous interconnected network and oxygen and water vapor permeability. Using bovine serum albumin as a model protein, we demonstrated that protein loading and release from the PBSA AMs were affected by the membrane structure and the presence of residual PA. In vitro studies showed highest antibiofilm efficiency both in inhibition and dispersion (up to 80%). Normalized in vitro cytotoxicity standard assays revealed that unloaded and DB-loaded PBSA membranes met cytocompatibility criteria required for wound dressing applications.

En France, le cancer du sein est le cancer le plus fréquent chez les femmes et la deuxième cause de décès par cancer. Si la mortalité relative du cancer du sein a diminué ces dernières années, de nombreux progrès restent à faire pour les 20 à 25 % de patientes qui évoluent vers un stade métastatique. Cette étude s’appuie sur la réalisation de 103 entretiens semi-directifs réalisés auprès d’acteurs de la prise en charge en oncologie et de patientes atteintes de cancer du sein métastatique. L’objectif de ce travail de thèse était d’étudier : l’impact des nouvelles thérapies ciblées sur l’organisation de la prise en charge et leur diffusion, le vécu des patientes avec les traitements et leur adhésion thérapeutique. Nous avons montré, dans un premier temps, que les nouvelles thérapies ciblées orales modifient la prise en charge traditionnelle des patientes qui restent moins dans la sphère d’influence de la structure de soins et qu’il existe de nombreux freins à la diffusion de ces traitements chez les oncologues. Dans un deuxième temps, nous avons mis en avant les représentations positives des patientes sur les thérapies ciblées orales qui sont sources d’espoir et améliorent la qualité de vie. Pour autant, le vécu au quotidien vient parfois nuancer ces représentations et peut révéler des freins à l’adhésion thérapeutique tels que l’apparition d’effets secondaires ou les responsabilités engendrées par la demande d’autonomie. Enfin, dans un troisième temps, nous avons mis en exergue le manque de connaissances des patientes sur les innovations thérapeutiques et l’importance de la décision partagée avec l’oncologue tout particulièrement face à la chronicisation progressive de certains cancers du sein métastatiques. / In France, breast cancer is the most common cancer amongst women and the second cause of cancer death. Although breast cancer relative mortality has decreased significantly in recent years, there is still a lot to be done for the remaining 20% to 25% of patients who have metastasized. This study is based on 103 semi-structured interviews carried out with oncology healthcare professionals and metastatic breast cancer patients. The aim of this work was to study the impact that new targeted therapies can have on managing the patient’s care and their diffusion, the patients’ experiences with these treatments, their social representations and their therapeutic adherence. First, we showed that new oral targeted therapies disrupt the traditional steps in a patient care. Patients arefar less in the hospital’s sphere of influence and there are many barriers to the dissemination of these treatments for oncologists. Second, we highlighted the positive representations of oral targeted therapies that patients have, which is bringing them hope and improving their quality of life. However, the day-to-day experience with the treatment sometimes moderates these representations and may reveal obstacles to their therapeutic adherence such as side effects or the responsibilities created by the demand for autonomy. Finally, we showed the lack of patients’ knowledge about therapeutic innovations and the importance of the implementation of shared decision making with the oncologist, especially in the face of the progressive chronicisation of a number of metastatic breast cancers.

<p>Lung cancer is the third most prevalent cancer in the world; however it is the leading cause of cancer related deaths worldwide. Non-small cell lung cancer (NSCLC) accounts for ~85% of the lung cancer cases. The current strategies to treat NSCLC patients with frequent causal genetic mutations is through targeted therapeutics. Approximately 10-35% of NSCLC patient tumors have activated mutations in the Epidermal Growth Factor Receptor (EGFR) resulting in uncontrolled cellular proliferation. The standard-of care for such patients is EGFR-Tyrosine Kinase Inhibitors (EGFR-TKIs), a class of targeted therapeutics that specifically inhibit EGFR activity. One such EGFR-TKI used in this study is erlotinib. Following erlotinib treatment, tumors rapidly regress at first; however, over 50% of patients develop erlotinib resistance within a year post treatment. Development of resistance remains to be the major challenge in treatment of NSCLC using EGFR-TKIs such as erlotinib. </p> <p>In approximately 60% of cases, acquired erlotinib resistance in patients is attributed to a secondary mutation in EGFR, whereas in about 20% of cases, activation of alternative signaling pathways is the reported mechanism. For the remaining 15-20% of <a>cases</a> the mechanism of resistance remains unknown. Therefore, it can be speculated that the common methods used to identify genetic mutations in tumors post erlotinib treatment, such as histologic analysis and genetic screening may fail to identify alterations in epigenetic mediators of erlotinib resistance, also including microRNAs (miRNAs). MiRNAs are short non-coding RNAs that post-transcriptionally negatively regulate their target transcripts. Hence, in this study two comprehensive screens were simultaneously conducted in erlotinib sensitive cells: 1) a genome-wide knock-out screen, conducted with the hypothesis that loss of function of certain genes drive erlotinib resistance, 2) a miRNA overexpression screen, conducted with the hypothesis that certain miRNAs drive the development of erlotinib resistance when overexpressed. The overreaching goal of the study was to identify novel drivers of erlotinib resistance such as microRNAs or other epigenetic factors in NSCLC.</p><p>The findings of this study led to the identification of a tumor suppressive protein and an epigenetic regulator, SUV420H2 (KMT5C) that has never been reported to be involved in erlotinib resistance. On the other hand, the miRNA overexpression screen identified five miRNAs that contribute to erlotinib resistance that were extensively analyzed using multiple bioinformatic tools. It was predicted that the miRNAs mediate erlotinib resistance via multiple pathways, owing to the ability of each miRNA to target multiple transcripts via partial complementarity. Importantly, a correlation between the two screens was identified clearly supporting the use of two simultaneous screens as a reliable technique to determine highly significant miRNA-target interactions. Overall, the findings from this study suggest that epigenetic factors, such as histone modifiers and miRNAs function as critical mediators of erlotinib resistance, possibly belonging to the 15-20% of NSCLC cases with unidentified mechanisms involved in erlotinib resistance.</p><p></p>