Random Insertion Research Articles

In Vivo Sampling of Intracellular Heterogeneity of Pseudomonas putida Enables Multiobjective Optimization of Genetic Devices.

The inner physicochemical heterogeneity of bacterial cells generates three-dimensional (3D)-dependent variations of resources for effective expression of given chromosomally located genes. This fact has been exploited for adjusting the most favorable parameters for implanting a complex device for optogenetic control of biofilm formation in the soil bacterium Pseudomonas putida. To this end, a DNA segment encoding a superactive variant of the Caulobacter crescendus diguanylate cyclase PleD expressed under the control of the cyanobacterial light-responsive CcaSR system was placed in a mini-Tn5 transposon vector and randomly inserted through the chromosome of wild-type and biofilm-deficient variants of P. putida lacking the wsp gene cluster. This operation delivered a collection of clones covering a whole range of biofilm-building capacities and dynamic ranges in response to green light. Since the phenotypic output of the device depends on a large number of parameters (multiple promoters, RNA stability, translational efficacy, metabolic precursors, protein folding, etc.), we argue that random chromosomal insertions enable sampling the intracellular milieu for an optimal set of resources that deliver a preset phenotypic specification. Results thus support the notion that the context dependency can be exploited as a tool for multiobjective optimization, rather than a foe to be suppressed in Synthetic Biology constructs.

Every little bit helps: A single-residue switch in a vascular AAV enables blood-brain barrier penetration

Every little bit helps: A single-residue switch in a vascular AAV enables blood-brain barrier penetration

Salmonella Enteritidis T1SS protein SiiD inhibits NLRP3 inflammasome activation via repressing the mtROS-ASC dependent pathway.

Inflammasome activation is an essential innate immune defense mechanism against Salmonella infections. Salmonella has developed multiple strategies to avoid or delay inflammasome activation, which may be required for long-term bacterial persistence. However, the mechanisms by which Salmonella evades host immune defenses are still not well understood. In this study, Salmonella Enteritidis (SE) random insertion transposon library was screened to identify the key factors that affect the inflammasome activation. The type I secretion system (T1SS) protein SiiD was demonstrated to repress the NLRP3 inflammasome activation during SE infection and was the first to reveal the antagonistic role of T1SS in the inflammasome pathway. SiiD was translocated into host cells and localized in the membrane fraction in a T1SS-dependent and partially T3SS-1-dependent way during SE infection. Subsequently, SiiD was demonstrated to significantly suppress the generation of mitochondrial reactive oxygen species (mtROS), thus repressing ASC oligomerization to form pyroptosomes, and impairing the NLRP3 dependent Caspase-1 activation and IL-1β secretion. Importantly, SiiD-deficient SE induced stronger gut inflammation in mice and displayed NLRP3-dependent attenuation of the virulence. SiiD-mediated inhibition of NLRP3 inflammasome activation significantly contributed to SE colonization in the infected mice. This study links bacterial T1SS regulation of mtROS-ASC signaling to NLRP3 inflammasome activation and reveals the essential role of T1SS in evading host immune responses.

Building an improved transcription factor-centered yeast one hybrid system to identify DNA motifs bound by protein comprehensively

BackgroundIdentification of the motifs bound by a transcription factor (TF) is important to reveal the function of TF. Previously, we built a transcription factor centered yeast one hybrid (TF-Centered Y1H) that could identify the motifs bound by a target TF. However, that method was difficult to comprehensively identify all the motifs bound by a TF.ResultsHere, we build an improved TF-Centered Y1H to comprehensively determine the motifs bound by a target TF. Recombination-mediated cloning in yeast was performed to construct a saturated prey library that contains 7 random base insertions. After TF-Centered Y1H screening, all the positive clones were pooled together to isolate pHIS2 vector. The insertion regions of pHIS2 were PCR amplified and the PCR product was subjected to high-throughput sequencing. The insertion sequences were then retrieved and analyzed using MEME program to identify the potential motifs bound by the TF. Using this technology, we studied the motifs bound by an ethylene-responsive factor (BpERF2) from birch. In total, 22 conserved motifs were identified, and most of them are novel cis-acting elements. Both the yeast one hybrid and electrophoretic mobility shift assay verified that the obtained motifs could be bound by BpERF2. In addition, chromatin immunoprecipitation (ChIP) study further suggested that the identified motifs can be bound by BpERF2 in cells of birch. These results together suggested that this technology is reliable and has biological significance.ConclusionThis method will have wide application in DNA-protein interaction studies.

Efficient Transgenesis in African Turquoise Killifish Using a Gibson Assembly-Based Tol2 Transposon System.

The short-lived African killifish Nothobranchius furzeri is an attractive genetic model for vertebrate aging and regeneration studies. The utilization of genetically modified animals is a common strategy for unveiling molecular mechanisms responsible for a biological phenomenon. Here, we report a highly efficient protocol for generating transgenic African killifish using the Tol2 transposon system, which creates random insertions in the genome. Transgenic vectors carrying gene-expression cassettes of interest and an eye-specific marker for transgene identification can be quickly assembled through Gibson assembly. The development of this new pipeline will facilitate transgenic reporter assays and gene-expression-related manipulations in African killifish.

Rapid and precise genotyping of transgene zygosity in mice using an allele-specific method.

Precise determination of transgene zygosity is essential for use of transgenic mice in research. Because integration loci of transgenes are usually unknown due to their random insertion, assessment of transgene zygosity remains a challenge. Current zygosity genotyping methods (progeny testing, qPCR, and NGS-computational biology analysis) are time consuming, prone to error or technically challenging. Here, we developed a novel method to determine transgene zygosity requiring no knowledge of transgene insertion loci. This method applies allele-specific restriction enzyme digestion of PCR products (RE/PCR) to rapidly and reliably quantify transgene zygosity. We demonstrate the applicability of this method to three transgenic strains of mice (Atm TgC3001L, Nes-Cre, and Syn1-Cre) harboring a unique restriction enzyme site on either the transgene or its homologous sequence in the mouse genome. This method is as accurate as the gold standard of progeny testing but requires 2 d instead of a month or more. It is also exceedingly more accurate than the most commonly used approach of qPCR quantification. Our novel method represents a significant technical advance in determining transgene zygosities in mice.

Exploration of the skeletal phenotype of the Col1a1 +/Mov13 mouse model for haploinsufficient osteogenesis imperfecta type 1.

Osteogenesis Imperfecta is a rare genetic connective tissue disorder, characterized by skeletal dysplasia and fragile bones. Currently only two mouse models have been reported for haploinsufficient (HI) mild Osteogenesis Imperfecta (OI); the Col1a1 +/Mov13 (Mov13) and the Col1a1 +/-365 mouse model. The Mov13 mice were created by random insertion of the Mouse Moloney leukemia virus in the first intron of the Col1a1 gene, preventing the initiation of transcription. Since the development of the Mov13 mice almost four decades ago and its basic phenotypic characterization in the 90s, there have not been many further studies. We aimed to extensively characterize the Mov13 mouse model in order to critically evaluate its possible use for preclinical studies of HI OI. Bone tissue from ten heterozygous Mov13 and ten wild-type littermates (WT) C57BL/6J mice (50% males per group) was analyzed at eight weeks of age with bone histomorphometry, micro computed tomography (microCT), 3-point bending, gene expression of different collagens, as well as serum markers of bone turnover. The Mov13 mouse presented a lower bone strength and impaired material properties based on our results of 3-point bending and microCT analysis respectively. In contrast, no significant differences were found for all histomorphometric parameters. In addition, no significant differences in Col1a1 bone expression were present, but there was a significant lower P1NP concentration, a bone formation marker, measured in serum. Furthermore, bone tissue of Mov13 mice presented significantly higher expression of collagens (Col1a2, Col5a1 and Col5a2), and bone metabolism markers (Bglap, Fgf23, Smad7, Edn1 and Eln) compared to WT. Finally, we measured a significantly lower Col1a1 expression in heart and skin tissue and also determined a higher expression of other collagens in the heart tissue. Although we did not detect a significant reduction in Col1a1 expression in the bone tissue, a change in bone structure and reduction in bone strength was noted. Regrettably, the variability of the bone phenotype and the appearance of severe lymphoma in adult Mov13 mice, does not favor their use for the testing of new long-term drug studies. As such, a new HI OI type 1 mouse model is urgently needed.

Semantics-Preserving Reinforcement Learning Attack Against Graph Neural Networks for Malware Detection

As an increasing number of deep-learning-based malware scanners have been proposed, the existing evasion techniques, including code obfuscation and polymorphic malware, are found to be less effective. In this work, we propose a reinforcement learning based semantics-preserving (i.e. functionality-preserving) attack against black-box GNNs (Graph Neural Networks) for malware detection. The key factor of adversarial malware generation via semantic <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Nops</monospace> insertion is to select the appropriate semantic <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Nops</monospace> and their corresponding basic blocks. The proposed attack uses reinforcement learning to automatically make these “how to select” decisions. To evaluate the attack, we have trained two kinds of GNNs with three types (e.g., Backdoor, Trojan, and Virus) of Windows malware samples and various benign Windows programs. The evaluation results have shown that the proposed attack can achieve a significantly higher evasion rate than four baseline attacks, namely the binary diversification attack, the semantics-preserving random instruction insertion attack, the semantics-preserving accumulative instruction insertion attack, and the semantics-preserving gradient-based instruction insertion attack.

Three dimensional printable multi‐arms poly(CL‐co‐TOSUO) for resilient biodegradable elastomer

AbstractThe development of artificial organs and implants demands new materials with both processability and functionality. A series of star like poly(CL‐co‐TOSUO) copolymers with different molecular weights are synthesized via ring‐opening polymerization. The polymers are modified with norbornene end groups for rapid photo‐crosslinking. Due to the random insertion of 1,4,8‐trioxaspiro‐[4,6]‐9‐undecanone and well‐constructed permanent chemical network, the photo‐crosslinked elastomers are more flexible and resilient than conventional poly(ε‐caprolactone), with 84% resilience above the melting point. Their thermal and mechanical properties can be fine‐tuned by varying the molecular weights of the precursors. The polymer can be fabricated into complex shapes by digital light processing 3D printing. Combined with good cytocompatibility and degradability, the highly flexible elastomers show great potential for tissue engineering.

Fitness-Conditional Genes for Soil Adaptation in the Bioaugmentation Agent Pseudomonas veronii 1YdBTEX2.

Strain inoculation (bioaugmentation) is a potentially useful technology to provide microbiomes with new functionalities. However, there is limited understanding of the genetic factors contributing to successful establishment of inoculants. This work aimed to characterize the genes implicated in proliferation of the monoaromatic compound-degrading Pseudomonas veronii 1YdBTEX2 in nonsterile polluted soils. We generated two independent mutant libraries by random minitransposon-delivered marker insertion followed by deep sequencing (Tn-seq) with a total of 5.0 × 105 unique insertions. Libraries were grown in multiple successive cycles for up to 50 generations either in batch liquid medium or in two types of soil microcosms with different resident microbial content (sand or silt) in the presence of toluene. Analysis of gene insertion abundances at different time points (passed generations of metapopulation growth), in comparison to proportions at start and to in silico generated randomized insertion distributions, allowed to define ~800 essential genes common to both libraries and ~2,700 genes with conditional fitness effects in either liquid or soil (195 of which resulted in fitness gain). Conditional fitness genes largely overlapped among all growth conditions but affected approximately twice as many functions in liquid than in soil. This indicates soil to be a more promiscuous environment for mutant growth, probably because of additional nutrient availability. Commonly depleted genes covered a wide range of biological functions and metabolic pathways, such as inorganic ion transport, fatty acid metabolism, amino acid biosynthesis, or nucleotide and cofactor metabolism. Only sparse gene sets were uncovered whose insertion caused fitness decrease exclusive for soils, which were different between silt and sand. Despite detectable higher resident bacteria and potential protist predatory counts in silt, we were, therefore, unable to detect any immediately obvious candidate genes affecting P. veronii biological competitiveness. In contrast to liquid growth conditions, mutants inactivating flagella biosynthesis and motility consistently gained strong fitness advantage in soils and displayed higher growth rates than wild type. In conclusion, although many gene functions were found to be important for growth in soils, most of these are not specific as they affect growth in liquid minimal medium more in general. This indicates that P. veronii does not need major metabolic reprogramming for proliferation in soil with accessible carbon and generally favorable growth conditions. IMPORTANCE Restoring damaged microbiomes is still a formidable challenge. Classical widely adopted approaches consist of augmenting communities with pure or mixed cultures in the hope that these display their intended selected properties under in situ conditions. Ecological theory, however, dictates that introduction of a nonresident microbe is unlikely to lead to its successful proliferation in a foreign system such as a soil microbiome. In an effort to study this systematically, we used random transposon insertion scanning to identify genes and possibly, metabolic subsystems, that are crucial for growth and survival of a bacterial inoculant (Pseudomonas veronii) for targeted degradation of monoaromatic compounds in contaminated nonsterile soils. Our results indicate that although many gene functions are important for proliferation in soil, they are general factors for growth and not exclusive for soil. In other words, P. veronii is a generalist that is not a priori hindered by the soil for its proliferation and would make a good bioaugmentation candidate.

Enhancer trap lines with GFP driven by smad6b and frizzled1 regulatory sequences for the study of epithelial morphogenesis in the developing zebrafish inner ear.

Live imaging in the zebrafish embryo using tissue-specific expression of fluorescent proteins can yield important insights into the mechanisms that drive sensory organ morphogenesis and cell differentiation. Morphogenesis of the semicircular canal ducts of the vertebrate inner ear requires a complex rearrangement of epithelial cells, including outgrowth, adhesion, fusion and perforation of epithelial projections to generate pillars of tissue that form the hubs of each canal. We report the insertion sites and expression patterns of two enhancer trap lines in the developing zebrafish embryo, each of which highlight different aspects of epithelial cell morphogenesis in the inner ear. A membrane-linked EGFP driven by smad6b regulatory sequences is expressed throughout the otic epithelium, most strongly on the lateral side of the ear and in the sensory cristae. A second enhancer trap line, with cytoplasmic EGFP driven by frizzled1 (fzd1) regulatory sequences, specifically marks cells of the ventral projection and pillar in the developing ear, and marginal cells in the sensory cristae, together with variable expression in the retina and epiphysis, and neurons elsewhere in the developing central nervous system. We have used a combination of methods to identify the insertion sites of these two transgenes, which were generated through random insertion, and show that Targeted Locus Amplification is a rapid and reliable method for the identification of insertion sites of randomly inserted transgenes.

On the origin of the genetic code.

Mechanisms underlying how the genetic code was generated by Darwinian selection have remained elusive since the code was cracked in 1965. Here, I propose a hypothesis on the emergence of the genetic code and predict that its emergence was driven by sequential distinct selective pressures. According to the hypothesis, aminoacyl-RNAs for Glu, Asp, Lys, Tyr, His, Arg, Cys and Ser were first selected as cartridge-type subunits of three-subunit ribozymes. Aminoacyl-RNA subunits acting as cofactors were accommodated by the proto P-site of the large subunit of ribozymes. Importantly, I predict that there was no direct relationship between amino acids and codon and anticodon pairs. Duplication of the proto P-site could have created the proto A-site, enabling multi-subunit ribozymes to simultaneously interact with two-cartridge-type aminoacyl-RNA subunits. Random insertion of two cartridges would have instantly abolished enzymatic activity of multi-subunit ribozymes. On the other hand, if two tandemly aligned pairs of codons and anticodons specify two cartridges, dozens of different active pockets in multi-subunit ribozymes would have rapidly emerged, leading to the rise of extant organisms' metabolic pathways. The strong driving force of Darwinian selection described here could have created the primary genetic code for catalytic amino acids. Evolution of the protein translation system and events leading to the expansion of the genetic code until the time it was "frozen" are presented in detail.

Textual Augmentation Techniques Applied to Low Resource Machine Translation: Case of Swahili

In this work we investigate the impact of applying textual data augmentation tasks to low resource machine translation. There has been recent interest in investigating approaches for training systems for languages with limited resources and one popular approach is the use of data augmentation techniques. Data augmentation aims to increase the quantity of data that is available to train the system. In machine translation, majority of the language pairs around the world are considered low resource because they have little parallel data available and the quality of neural machine translation (NMT) systems depend a lot on the availability of sizable parallel corpora. We study and apply three simple data augmentation techniques popularly used in text classification tasks; synonym replacement, random insertion and contextual data augmentation and compare their performance with baseline neural machine translation for English-Swahili (En-Sw) datasets. We also present results in BLEU, ChrF and Meteor scores. Overall, the contextual data augmentation technique shows some improvements both in the EN -&gt; SW and SW -&gt; EN directions. We see that there is potential to use these methods in neural machine translation when more extensive experiments are done with diverse datasets.

HDR-based CRISPR/Cas9-mediated Knockout of PD-L1 in C57BL/6 Mice.

The immune-inhibitory molecule programmed cell death ligand 1 (PD-L1) has been shown to play a role in pathologies such as autoimmunity, infections, and cancer. The expression of PD-L1 not only on cancer cells but also on non-transformed host cells is known to be associated with cancer progression. Generation of PD-L1 deficiency in the murine system enables us to specifically study the role of PD-L1 in physiological processes and diseases. One of the most versatile and easy to use site-specific gene editing tools is the CRISPR/Cas9 system, which is based on an RNA-guided nuclease system. Similar to its predecessors, the Zinc finger nucleases or transcription activator-like effector nucleases (TALENs), CRISPR/Cas9 catalyzes double-strand DNA breaks, which can result in frameshift mutations due to random nucleotide insertions or deletions via non-homologous end joining (NHEJ). Furthermore, although less frequently, CRISPR/Cas9 can lead to insertion of defined sequences due to homology-directed repair (HDR) in the presence of a suitable template. Here, we describe a protocol for the knockout of PD-L1 in the murine C57BL/6 background using CRISPR/Cas9. Targeting of exon 3 coupled with the insertion of a HindIII restriction site leads to a premature stop codon and a loss-of-function phenotype. We describe the targeting strategy as well as founder screening, genotyping, and phenotyping. In comparison to NHEJ-based strategy, the presented approach results in a defined stop codon with comparable efficiency and timelines as NHEJ, generates convenient founder screening and genotyping options, and can be swiftly adapted to other targets.

A New Image Encryption Scheme Based on 6D Hyperchaotic System and Random Signal Insertion

A New Image Encryption Scheme Based on 6D Hyperchaotic System and Random Signal Insertion

Enterococcus faecalis NADH Peroxidase-Defective Mutants Stain Falsely in Colony Zymogram Assay for Extracellular Electron Transfer to Ferric Ions.

Enterococcus faecalis cells can reduce ferric ions and other electron acceptors by extracellular electron transfer (EET). To find mutants with enhanced or defective EET, strain OG1RF with random transposon insertions in the chromosome was screened for ferric reductase activity by colony zymogram staining using the chromogenic ferrous-chelating compound Ferrozine. The screen revealed npr, eetB, and ndh3 mutants. The aberrant ferric reductase phenotype of Npr (NADH peroxidase)-defective mutants was found to be a property of colonies and not apparent with washed cells grown in liquid culture. EetB- and Ndh3-defective mutants, in contrast, consistently showed low ferric reductase activity. It is concluded that colony zymogram staining for ferric reductase activity using Ferrozine can be misleading, especially through false negative results. It is suggested that hydrogen peroxide produced in the colony quenches the zymogram staining. In addition, it is demonstrated that the negative effect of heme on EET to ferric ion in E. faecalis is relieved by cytochrome bd deficiency. The findings can help to identify bacteria with EET ability and contribute to our understanding of EET in Gram-positive bacteria and the physiology of E. faecalis.

Beneficial and detrimental genes in the cellular response to replication arrest.

DNA replication is essential for all living organisms. Several events can disrupt replication, including DNA damage (e.g., pyrimidine dimers, crosslinking) and so-called "roadblocks" (e.g., DNA-binding proteins or transcription). Bacteria have several well-characterized mechanisms for repairing damaged DNA and then restoring functional replication forks. However, little is known about the repair of stalled or arrested replication forks in the absence of chemical alterations to DNA. Using a library of random transposon insertions in Bacillus subtilis, we identified 35 genes that affect the ability of cells to survive exposure to an inhibitor that arrests replication elongation, but does not cause chemical alteration of the DNA. Genes identified include those involved in iron-sulfur homeostasis, cell envelope biogenesis, and DNA repair and recombination. In B. subtilis, and many bacteria, two nucleases (AddAB and RecJ) are involved in early steps in repairing replication forks arrested by chemical damage to DNA and loss of either nuclease causes increased sensitivity to DNA damaging agents. These nucleases resect DNA ends, leading to assembly of the recombinase RecA onto the single-stranded DNA. Notably, we found that disruption of recJ increased survival of cells following replication arrest, indicating that in the absence of chemical damage to DNA, RecJ is detrimental to survival. In contrast, and as expected, disruption of addA decreased survival of cells following replication arrest, indicating that AddA promotes survival. The different phenotypes of addA and recJ mutants appeared to be due to differences in assembly of RecA onto DNA. RecJ appeared to promote too much assembly of RecA filaments. Our results indicate that in the absence of chemical damage to DNA, RecA is dispensable for cells to survive replication arrest and that the stable RecA nucleofilaments favored by the RecJ pathway may lead to cell death by preventing proper processing of the arrested replication fork.

A transposon system for random insertion of a gene expression cassette into the chromosome of Bacillus subtilis

Bacillus subtilis is a robust industrial workhorse for the production of heterologous proteins. Chromosomal integration-based protein production has advantages over plasmid-based methods. Considering that the expression level of a gene is affected by its location in the chromosome, it is important to find an optimal integration site for the gene to be expressed. This work establishes a method for random insertion of a gene expression cassette into chromosomes, enabling the screening of optimal integration sites for high-level protein production. Specifically, a gene expression cassette and a chloromycetin-resistance marker are assembled into a transposon. This transposon is inserted between the promoter and the ribosomal binding site of the zeocin-resistance marker in the chromosome, which blocks the transcription of the zeocin-resistance gene. Transposase Himar1-mediated transposition of this transposon activates the zeocin-resistance marker, which can be selected on plates containing both chloromycetin and zeocin. The transposition frequency was over 10−5. This method was used to select proper insertion sites for the expression cassette of methyl parathion hydrolase (MPH). Compared with the common integration site amyE, the expression level of MPH was increased up to 50 % at the yjbH site.

Genome editing is induced in a binary manner in single human cells.

Even when precise nucleotide manipulations are intended, the outcomes of genome editing can be diverse, often including random insertions and deletions. The combinations and frequencies of these different outcomes in single cells are critical not only in the generation of genetically modified cell lines but also in the evaluation of the clinical effects of genome editing therapies. However, current methods only analyze cell populations, not single cells. Here, we utilized the Single Particle isolation System (SPiS) for the efficient isolation of single cells to systematically analyze genome editing results in individual human cultured cells. As a result, we discovered that genome editing induction has a binary nature, that is, the targetalleles of cells tend to be all edited or not edited at all. This study enhances our understanding of the induction pattern of genome editing and provides a new strategy to analyze genome editing outcomes in single cells.

Mutation Spectrum of FLT3 and Significance of Non-Canonical FLT3 Mutations in Hematological Malignancies

Background FLT3 mutations are detected in 15%-35% of newly diagnosed AML patients. The canonical FLT3 mutation types are internal tandem repeats (ITD) in the JMD and TKD1 regions and point mutations at codons D835/I836 (TKD). Conventional sequencing for FLT3 mutations only covers the above-mentioned hotspot regions. With the advent of next generation sequencing which spans the full-length FLT3, a series of non-canonical mutations have also been reported. However, the kinase activities and clinical characteristics of these mutations remain unclear. Methods To profile the spectrum of FLT3 mutations, we retrospectively analyzed 869 consecutive newly diagnosed AML, ALL, and MDS patients with FLT3 mutation by NGS-seq from June 2016 to December 2020 at Jiangsu Institute of Hematology. The clinical data of 316 AML (excluding APL) patients were retrospectively collected and analyzed after informed consent. Since previous studies have elaborated the FLT3 functional characteristics of FLT3-ITD and FLT3 point mutations, we focus on other FLT3 mutation types in this study. Seven representative FLT3 mutations, including four deletion mutations in JMD, one frameshift mutation in JMD, one deletion mutation in TKD1 and one ITD in TKD2, were constructed to study the kinase activities and sensitivities to FLT3 inhibitors in vitro. Results FLT3 mutational spectrums of 869 patients showed that the NGS method could detect numerous novel non-canonical FLT3 mutations, accounting for 21.2% of FLT3 mutation events compared with conventional tests (Figure 1). According to the changes in gene structures resulted by different mutations, we classified the non-canonical mutations into four types: non-canonical point mutations (NCPM) (19.2%), deletion (0.7%), frameshift (0.8%), and ITD outside the JMD and TKD1 regions (0.5%). Given the enormous differences in frequency, NCPM was further divided into FLT3 recurrent point mutations of NPM (FLT3-NCPM-RPM) and FLT3 non-recurrent point mutations (FLT3-NRPM). Kaplan-Meier analyses of the RFS and OS showed that there is no difference between canonical FLT3-TKD and FLT3-NCPM-RPM while statistical differences were observed not only between canonical FLT3-TKD and FLT3-NRPM but also between FLT3-NCPM-RPM and FLT3-NRPM. Due to the similarities in prognosis and kinase activity, classical FLT3-TKD and FLT3-NCPM-RPM were grouped into a class called FLT3 recurrent point mutations (FLT3-RPM). Further prognosis analyses indicated that patients with FLT3-ITD, FLT3-RPM and FLT3-NRPM had significant differences in RFS (P=0.005) and OS (P=0.01). Interestingly, the presence of FLT3-NRPM may confer a favorable prognosis compared with FLT3-ITD and FLT3-RPM. Multivariable analysis using Cox models also confirmed that FLT3-NRPM had a survival benefit over FLT3-ITD and FLT3-RPM. The significant difference in the prognosis of patients with three FLT3 mutation subtypes and Cox models proved the rationality of this strategy. Besides the non-canonical point mutation, a series of FLT3 mutations with small deletions and frameshift alterations were also detected. In addition, we also discovered several completely random insertion mutations in the non-ITD region. According to our previous study, these mutageneses were considered a subtype of FLT3-ITD. Based on the results, FLT3 alterations could be divided into FLT3-ITD, FLT3-RPM, FLT3-NRPM, FLT3-deletion, and FLT3-frameshift groups. In vitro analysis showed that the deletion mutants in TKD1 (p.K614_G617del) and the FLT3-ITD mutant in TKD2 (p.S941_F942insRVS) had significantly higher kinase activity than wild-type FLT3. Whereas the deletion mutants in JMD have comparable phosphorylation levels to wild-type FLT3 (Figure 2). All tested deletion mutations and ITD in the TKD2 region were sensitive to both AC220 and sorafenib. Conclusion Conventional sequencing that only covers hotspot regions omitted part of FLT3 mutation and full-length sequencing of the FLT3 gene may be indispensable. FLT3 alterations detected by NGS-seq could be reclassified into 5 types: FLT3-ITD, FLT3-RPM, FLT3-NRPM, FLT3-deletion, and FLT3-frameshift. In vitro studies indicated patients with FLT3 deletion mutations and FLT3-ITD outside the JMD and TKD1 regions might be ought to be treated with FLT3 inhibitors. Our new classification of FLT3 mutation was expected to promote more precise management of patients with FLT3 non-canonical mutations. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal