Clone Detection Research Articles

Historical perspective of code clone refactorings in evolving software.

Cloning in software is generally perceived as a threat to its maintenance and that is why it needs to be managed properly. Understanding clones from a historical perspective is essential for effective clone management. Analysis of code refactorings performed on clones in previous releases will help developers in taking decisions about clone refactoring in future releases. In this paper we perform a longitudinal study on the evolution of clone refactorings in various versions of five software systems. To perform a systematic study on clone refactoring evolution, we define clone evolution patterns for studying refactorings in a formal notation. Our results show that only a small proportion of code clones are refactored between the versions and most of the refactorings are inconsistent within clone classes. Moreover, clone refactorings may cause clone removal. Analysis of the source code of refactored clones reveals similar reasons of inconsistent refactorings and clone removal for five Java systems. This analysis will help in devising appropriate strategies for managing clone refactorings in software and hence provide foundation for devising better clone management tools.

Substantial uneven proliferation of CD4+ T cells during recovery from acute HIV infection is sufficient to explain the observed expanded clones in the HIV reservoir

The HIV reservoir is a population of 1–10 million anatomically dispersed, latently infected memory CD4+ T cells in which HIV DNA is quiescently integrated into human chromosomal DNA. When antiretroviral therapy (ART) is stopped and HIV replication initiates in one of these cells, systemic viral spread resumes, rekindling progression to AIDS. Therefore, HIV latency prevents cure. The detection of many populations of identical HIV sequences at unique integration sites implicates CD4+ T cell proliferation as the critical driver of reservoir sustainment after a prolonged period of effective ART. Initial reservoir formation occurs during the first week of primary infection usually before ART is started. While empirical data indicates that both de novo infection and cellular proliferation generate latently infected cells during early untreated infection, it is not known which of these mechanisms is predominant. We developed a mathematical model that recapitulates the profound depletion and brisk recovery of CD4+ T cells, reservoir creation, and viral load trajectory during primary HIV infection. We extended the model to stochastically simulate individual HIV reservoir clones. This model predicts the first detection of HIV infected clones approximately 5 weeks after infection as has recently been shown in vivo and suggests that substantial, uneven proliferation among clones during the recovery from CD4+ lymphopenia is the most plausible explanation for the observed clonal reservoir distribution during the first year of infection.

Software code refactoring based on deep neural network‐based fitness function

SummaryRefactoring is extensively recognized for enhancing the internal structure of object‐oriented software while preserving its external behavior. However, determining refactoring opportunities is challenging for designers and researchers alike. In recent years, machine learning algorithms have shown a great possibility of resolving this issue. This study proposes a deep neural network‐based fitness function (DNNFF) to resolve the software refactoring issue. This study suggests an effective learning technique that automatically featured extracted from the trained models and predicted code clones to recommend which category to refactor. The software engineers automatically assess the recommended refactoring solutions using Genetic Algorithms (GA) for minimum iterations. A Deep Neural Networks (DNN) utilizes these training instances to assess the refactoring solutions for the residual iterations. The refactoring process primarily depends on software designers' skills and perceptions. The simulation findings demonstrate that the suggested DNNFF model enhances the code change score of 98.7%, automatic refactoring score of 97.3%, defect correlation ratio of 96.9%, refactoring precision ratio of 95.9%, flaw detection ratio of 94.4%, and reduces the execution time of 10.2% compared to other existing methods.

VULDEFF: Vulnerability detection method based on function fingerprints and code differences

The significant increase in Open Source Software has led to a sharp increase in code cloning vulnerabilities. Code similarity detection methods are usually used to detect these vulnerabilities. However, cloned code often modifies the original code to varying degrees, and the difference between vulnerable code and patch code can be very small. Traditional code similarity detection methods cannot effectively detect common mutation patterns in code cloning and distinguish vulnerable code from patch code. The paper proposes a vulnerability detection method named VULDEFF based on function fingerprints and code differences. This paper designs a lightweight function fingerprint method based on the Context Triggered Piecewise Hashing algorithm, which can characterize the basic syntax features of function source code. In particular, the fingerprint of the vulnerable function contains the syntax features, vulnerability features, and patch features of the vulnerable function. VULDEFF detects whether target function has vulnerabilities by searching target function fingerprint in the vulnerable function fingerprint library. Compared with five advanced software vulnerability detection tools, VULDEFF significantly reduces the false positive and false negative rates while ensuring the scalability of vulnerability detection. VULDEFF discovered 111 new vulnerabilities in 10 open source projects.

Pro-Active 4SCAR2.0 Therapy with NGS-MRD Disease Monitoring in High-Risk B Cell Leukemia Management

BACKGROUND: Allogeneic hematopoietic stem cell transplantation (HCT) is a standard treatment for high-risk relapsed/refractory B-cell acute lymphoblastic leukemia (B-ALL). However, about 30-40% of patients still relapse after HCT. Chimeric antigen receptor-modified T cell (CAR-T) therapy holds great promise as a salvage approach. Although CD19-targeted CAR-Ts have demonstrated a rapid and significant anti-tumor effect in patients with B-cell malignancies, many still suffer disease relapse due to the decrease or loss of CD19 antigen. Therefore, novel curative approaches are needed. PURPOSE: The aim of this study is to determine the best timing for applying CAR-T therapy in high-risk B-ALL patients based on high-sensitivity analysis of tumor cells by next-generation sequencing (NGS) detecting minimal residual disease (MRD). An innovative CAR-T therapy concept, CAR2.0, has been proposed, which uses multi-antigen-targeted CAR-Ts to prevent antigen escape and prolong remission. Based on NGS-MRD detection, patients with potential tumor relapse can be identified early and managed accordingly. On the other hand, if extended negative NGS-MRD is found, the patient is deemed at low risk and no further treatment is needed. The study proposes to treat high-risk B-ALL patients in the first remission period with CAR2.0 followed by continued disease monitoring by NGS-MRD to avoid HCT. METHODS: B-ALL patients who have high tumor burden and high-risk genetic indication is enrolled with the understanding that CAR-T therapy will be applied in the first remission and disease will be continuously monitored by NGS-MRD. The tumor cells from initial diagnosis will be analyzed for target antigens by immunostaining and to detect clonally rearranged immunoglobulin gene sequences by NGS-MRD with sensitivity ≤10-6. Patients will receive 4SCAR19 (CD19scFv-CD28/CD27/CD3ζ-iCasp9) therapy in the first remission and will be followed up for CAR-Ts, B cells and IgG/IgA/IgM in the blood, and bi-monthly NGS-MRD for residual tumor cells in the bone marrow (BM). A booster CAR-T treatment targeting additional tumor antigen such as CD22 will be applied 2 months later. A consolidation CAR-T treatment based on known target antigens will be applied 6 months after the first CAR-T infusion. The cyclophosphamide/fludarabine chemo-conditioning will be included prior to the first CAR-T treatment only. The CAR-T infusion dose is 1-5x106 CAR-Ts/kg. The quality of apheresis cells, efficiency of gene transfer, T cell proliferation, CAR-T infusion dose and blood CAR copies are quantitatively documented. The end point for this study is the establishment of NGS-MRD negativity for 1 year, with continued follow-up for 5 years. RESULTS: A 20 years old female was diagnosed with BCR/ABL1 P190 positive B-ALL, with 82.52% tumor blasts in the BM. NGS identified two high-frequency IgH clones at 66.29% and 30.41%. The patient received induction chemotherapy (VTCP, CAM, VP+MTX) combined with TKI therapy (Nilotinib) and achieved complete remission (CR) after the first round of treatment based on both flow cytometry and NGS MRD analyses. During the fourth round of chemotherapy (Methotrexate-Cytarabine-Dexamethasone), the patient suffered severe oral and intestinal mucositis, along with kidney and liver toxicity and required week-long dialysis. A decision was made to stop chemotherapy and start the first 4SCAR19 therapy of 3.53x106 CAR-Ts/kg, while maintaining the TKI therapy. The CAR-T cells in blood, the B cell count and Ig levels were monitored, and a booster CD22/CD19 dual CAR-Ts (4SCAR2219) was given at a dose of 3.35 x106 CAR-Ts/kg on day 96 (see figure). Consistent with 4SCART safety profile, the patient experienced no greater than grade 1 CRS. NGS-MRD detection of tumor cell clones remained negative on day 96 after the first CAR-T infusion and at 6 months into the remission. In summary, the 4SCAR2.0 therapy regimen appeared safe and effective in the innovative NGS-MRD/4SCAR2.0 B-ALL study regimen with in vivo persistency for more than three months. The simultaneous expansion of CD19 mono-specific and CD22/CD19 bi-specific CAR-Ts supports a stable maintenance treatment with good life quality as compared with conventional chemotherapy. CONCLUSIONS: The combination of 4SCAR2.0 with NGS-MRD detection of tumor clones in the management of high-risk B-ALL appears to be safe and may be a logical option for patients not elected for HCT therapy. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Utilizing TCR-Seq to Detect Tabelecleucel, an Allogeneic Epstein-Barr Virus (EBV)-Specific T-Cell Immunotherapy, Post-Infusion

Background: Tabelecleucel is an off-the-shelf, allogeneic EBV-specific T-cell immunotherapy that targets EBV-antigen presenting cells in an HLA restricted manner. It is not genetically modified and therefore lacks a distinct universal molecular moiety to detect it post-infusion. To date, the most effective measure of EBV reactive T-cells in patients (pts) treated with tabelecleucel is the EBV-cytotoxic T lymphocyte precursor (CTLp) assay, which measures the number of EBV reactive CTL precursors in each sample but cannot distinguish between the pt's endogenous CTLs or the infused product CTLs. We have previously shown that EBV-CTLp peak correlates with pt response. Prior methods to identify circulating product CTLs have relied on differences in the genomic profiles of the pt and dosing product lot. However, the limit of detection for these assays to date has not been sensitive enough to reliably detect product CTLs post-infusion. One aspect of T-cell biology that could be utilized to track product specific CTL clones is the unique T cell receptors (TCR) which impart antigen specificity. TCRs are highly diverse and only a small minority of TCR clones are found to be shared across unrelated individuals. Therefore, we leveraged TCRβ sequencing of the CDR3 region to barcode unique TCR clones in tabelecleucel product lots and pt derived samples to detect product T cell clones post-infusion, and annotate likely EBV-targeting T cell clones based upon sequence homology to previously reported EBV-targeting TCR clones. Using this approach, we aimed to identify corollaries associated with clinical outcomes. Methods: Peripheral blood mononuclear cell samples from pts enrolled in the ALLELE (NCT03394365) study were obtained pre-dose on cycle (C) 1 day (D)1/D8/D15, C2D1/D8/D15 and at safety follow-up. EBV-CTLp was used to measure circulating EBV reactive T-cells, and both the pre-infusion baseline timepoint, and the timepoint of peak on-treatment EBV-CTL expansion were selected for downstream TCRβ sequencing. TCR repertoires for pt samples and the tabelecleucel dosing lot were assessed using TCRβ immunoSEQ assay. The captured TCR repertoires of pt samples were compared to the TCR repertoire of the dosing tabelecleucel lot to annotate product TCR clones and were compared against publicly available databases of previously identified EBV-specific TCRβ sequences to annotate putative EBV targeting TCR clones. Results: Analyses for an initial subset of 7 pts with EBV-positive post-transplant lymphoproliferative disease have been completed and are described here. Six unique tabelecleucel lots were used to treat these pts and were included in the analysis. On average 0.15 ± 0.04% of TCRs detected were shared between multiple lots, allowing the ability to leverage TCRβ sequences as a unique barcode to identify specific product lots. All 7 pts tested had TCRβ clones matching previously identified EBV-specific TCRs detected at baseline (median [min, max]: 3498 [1855, 12486]) and peak EBV-CTLp; clones detected at peak EBV-CTLp were either those previously detected at baseline (532 [129,1627]) or newly emerged (3622, [1801,5347]). TCRβ clones matching the dosing lot were identified in 6 out of 7 pts. Three pts had TCRβ clones matching the dosing lot present at baseline prior to tabelecleucel infusion (2 [1,5]). Post-infusion emergence of TCRs matching the dosing lot was detected in 5 pts (1 [1,3]). TCRβ sequencing on additional pts is currently underway. Association of post-infusion detection of dosing lot specific TCR clones with pt best overall response is ongoing and will be reported at the time of presentation. Conclusions: Utilizing TCRβ sequencing we can identify putative EBV-targeting TCR clones and dosing product TCR clones in pt TCR repertoires. Overall, an average of 12.5 ± 1.6% of captured TCR repertoires were annotated as EBV-targeting TCR clones. As publicly available reference datasets for EBV-specific TCRs are currently limited, this is likely an underrepresentation of the totality of EBV-specific TCRs present within the samples. Additionally, an average of 75.8% of the dosing product TCR clones detected were found exclusively after tabelecleucel infusion. Altogether, these results confirm that tracking TCRβ clones can be utilized as a novel assessment to detect post-infusion non-genetically modified allogenic cell therapies and future associations with clinical efficacy will be determined.

Spatial Organization of Hematopoietic Clones in the Bone Marrow of a Patient with Polycythemia Vera

All humans become genetic mosaics over time as our stem cells acquire somatic mutations throughout life. In the hematopoietic compartment, this phenomenon is termed clonal hematopoiesis (Steensma et al. 2015). Clonal hematopoiesis (CH) increases in prevalence with age and is associated with an increased risk for hematologic malignancy (Genovese et al. 2014; Jaiswal et al. 2014; Xie et al. 2014). CH harboring JAK2 mutations are frequently observed in healthy individuals; and yet JAK2 mutations are also the most frequent drivers of myeloproliferative neoplasms (MPNs) such as polycythemia vera (PV). MPNs are clonal diseases marked by dysregulated hematopoiesis and can progress to a secondary myelodysplastic syndrome or acute leukemia. There is increasing evidence that the clonal mutations driving MPNs arise early in life-potentially in utero-often years before the development of symptoms (Williams et al. 2022). While the intrinsic factors driving clonal expansion are increasingly known (Watson et al. 2020), the extrinsic factors driving the expansion of clones harboring these MPN driver mutations are not well characterized. Prior work has shown that clonal mutations are geographically constrained in solid organs such as skin, colon and esophagus (Lee-Six et al. 2018; Martincorena et al. 2015, 2018). Strikingly, in the esophagus non-malignant NOTCH1 mutant clones can even outcompete nascent malignant clones (Colom et al. 2021). In the hematopoietic compartment, the spatial organization of clones harboring these mutations and their interaction with the niche and each other is largely unknown. We hypothesized that clonal mutations are geographically constrained in the bone marrow niche as is seen in solid organs. To answer these questions, we developed methods for spatially-aware clonal mutation detection. As proof of principle, we characterized the hematopoietic clonal organization within a femur head explant from a 50 year-old man with PV undergoing hip arthroplasty. Briefly, the femur head was sectioned into 2mm slices, which were then further partitioned into approximately 300 bone marrow fragments (2mm3). Approximately 20% of the fragments were snap frozen and pulverized for DNA extraction. The remainder were formalin fixed and decalcified for correlative imaging. Targeted error-corrected sequencing was performed using a myeloid-specific panel (Duncavage et al. 2018) on 24 bone marrow fragments, and bulk peripheral blood and bone marrow aspirate. In the peripheral blood, JAK2 V617F (0.24 VAF) and TET2 L1081* (0.02 VAF) mutations were identified. In the bone marrow aspirate, the same JAK2 (0.62 VAF) and TET2 (0.05 VAF) mutations were identified in addition to a DNMT3A R882C mutation (0.02 VAF). Interestingly, peripheral blood sequencing two years prior identified the JAK2 and DNMT3A mutations, but not the TET2 mutation, suggesting recent clonal expansion of TET2 mutant clone. From the 24 sequenced bone marrow fragments, the geographical organization of clonal mutations in the bone marrow was determined (Figure 1). The JAK2 V617F clone was identified in all bone marrow fragments sequenced (0.12-0.69 VAF). The TET2 L1081* clone was identified in 11/24 bone marrow fragments with some regional constraint (0.03-0.05 VAF). The DNMT3A R882C mutation was detected in only a single bone marrow fragment (0.02 VAF) and co-localized with the JAK2 V617F and TET2 L1081* mutations. A previously undetected clonal SETBP1 E616V mutation was identified in two adjacent bone marrow fragments from sequential slices of the femur head (0.05 VAF and 0.10 VAF). Together this provides the first characterization of spatial organization of clonal mutations in the bone marrow niche for an MPN patient. Currently, we are utilizing digital droplet PCR targeting these detected mutations in ~60 additional bone marrow fragments to accurately quantify clone size across the tissue. We plan to employ multiplex fluorescent imaging of bone marrow fragments adjacent to detected clusters of mutations to identify the surrounding bone marrow niche elements in those regions. These findings suggest that bone marrow biopsy from a small section of bone marrow may not accurately describe disease state and clonal mutation burden in myeloid neoplasms. Future applications of these tools will help describe the extrinsic factors and local microenvironment that drive clonal evolution over time in the bone marrow. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Biologic Relevance of Variable Immunoglobulin Gene Diversity in B-Lymphoblastic Leukemia

Background: It remains unclear whether the variable biologic features of B-lymphoblastic leukemia (B-ALL) reflect distinct cells of origin in the hierarchy of early B cell development. Recombination activating genes (RAG) rearrange the immunoglobulin heavy chain (IgH) locus in B lymphoid progenitors, resulting in unique variable (Vh), diversity (D), and joining (Jh) gene rearrangements. In B-ALL, dominant IgH sequences are used for clone detection and surveillance. Further, subclone diversity resulting from RAG-mediated Vh gene switching is a prevalent feature of many B-ALLs, but its biologic basis has not been determined. To test whether discrete properties of each B-ALL reflect derivation from distinct B cell stages, we assessed the relationship between IgH variable gene (IGHV) characteristics and gene expression. Methods: We used targeted IgH sequencing (Illumina MiSeq) to identify dominant IgH clones and to assess the extent of IgH subclone diversification in pre-treatment specimens from 22 patients with B-ALL. We compared the gene expression profile (GEP) of cases with Vh-switch-mediated subclone diversity to those without. We then used single-cell RNA sequencing (scRNAseq) of 5 patients' B-ALL to test whether the capacity to undergo IgH diversification is patient-specific or clone-intrinsic, and we characterized variation in GEP between intrapatient IgH clones. Results: We discovered that each B-ALL with at least 1 dominant IgH clone demonstrates either minimal or extensive RAG-mediated subclone evolution. Gene set enrichment analysis (GSEA) revealed that genes associated with distinct B cell developmental stages distinguish these 2 groups. We detected dominant IgH clones in 19 of 22 patients (86.4%). In patients with 'diverse' B-ALL characterized by RAG-mediated intra-clonal diversification (N=9; 40.9%), we observed extensive subclone evolution arising from Vh recombination (median: 482 subclones; range: 2-2600) and enriched expression of genes associated with regulation of a hematopoietic stem cell state. In those without diversity (N=10; 45.5%) - the 'homogeneous' cohort - leukemia cells showed enriched expression of genes involving cell fate commitment. GSEA further revealed biologically distinct phenotypes between groups, as reflected in enrichment in the diverse cohort of Hallmark gene sets involving key cellular metabolic pathways including oxidative phosphorylation and glycolysis. Furthermore, distinct dominant IgH clones from a single patient could be individually diverse or homogenous with respect to Vh-switch-mediated subclone evolution. scRNAseq revealed that dominant IgH clones inconsistently express IGHV mRNA and differ from one another in their gene expression. In 4 of 5 patients, we detected IgH complementarity-determining region 3 (CDR3) mRNA expression from at least 1 clone. The GEP of IGHV-expressing leukemia cells was clearly resolved from other B-ALL cells in 3 of these 4 cases, with distinctive gene expression characterized by RAG1 pathway up-expression. Conclusions: Tumor evolution requires genetic diversity and is a fundamental property of cancer. Using primary B-ALL specimens, we assessed RAG-mediated intraclonal IgH variation to characterize the extent of diversity in B-ALL and to discover how it relates to fundamental biologic features of this common pediatric malignancy. We found that B-ALL may be divided into 2 groups: leukemias with RAG-mediated intra-clonal IGHV sequence diversification ('diverse') and those lacking diversity ('homogeneous'). GEP suggests that the difference in diversity of the IGHV repertoire reflects the B cell stage at leukemic initiation: B-ALL specimens with subclone diversity had a GEP suggestive of an earlier differentiation state compared to specimens with a homogenous IGHV repertoire. Furthermore, we discovered that a single leukemia can be simultaneously composed of both genetically stable and genetically diverse clonal populations. Our single-cell GEP data reveal that: 1) even within a single patient's leukemia, genetically-distinct clonal populations demonstrate a distinct profile of mRNA expression, and 2) B-ALL cells which express IGHV mRNA are characterized by greater expression of RAG pathway-related genes. Our findings suggest that variability in IGHV diversification in B-ALL is a useful marker of biological properties related to the B cell stage of origin.

Rare PNH Cells May be Considered As Genuine PNH Clone: Insights from Analysis of the 5-Year French Nation-Wide Multicenter Observational Study

Introduction The successive International Guidelines (Borowitz. et al. Cytometry B Clin Cytom 2010; Illingworth. et al. Cytometry B Clin Cytom 2018) recommend both detection and follow-up of rare PNH cells using high-resolution flow cytometry (FCM) analysis. These rare PNH cells, defined as GPI-deficient cells ranging from 0.01% to <0.1%, are distinguished from minor PNH clone, ranging from 0.1% to 1%. However, clinical relevance as well as long-term evolution of such rare PNH cells are still unknown. Objective We aimed to compare clinical and biological presentation of patients with rare PNH cells or a minor PNH clone along with long-term follow-up using data provided by the 5-year French nation-wide multicenter observational study, the Observatory of PNH Clones. Methods All patients of any age with a newly or previously diagnosed PNH clone or GPI-deficient cells ≥0.01%, detected in France by FCM from February 29th, 2016 to February 28th, 2021 were eligible for inclusion in this Observatory, provided that the center had participated in the inter-laboratory comparison program ongoing since 2013 (Debliquis et. al. Am J Clin Pathol 2015). For each patient, the baseline assessment was always considered as the initial PNH clone detection, even if it occurred before the initiation of the Observatory. Thus, this strategy allowed the collection of cases with long-term follow-up. A referent cytometrist at each center included patients in the e-CRF available on the CytHem-HPN website providing clinical and biological information as well as FCM raw data files. This study was approved by the national research ethics board. Results As of February 28th, 2021, 41 registered flow cytometry laboratories across France enrolled 637 patients with a PNH clone or GPI-deficient cells ≥ 0.01%. Selection of patients displaying rare PNH cells (i.e., o.o1% to <0.1%) or a minor PNH clone (i.e., 0.1% to 1%) on neutrophils led to the re-examination of 163 patient FCM files (provided by 26 centers) by the two principal investigators. FCM data and the e-CRF filling were both thoroughly reviewed leading to the update of roughly half of the submitted files. This enabled the validation of 138 patients at diagnosis and the exclusion of 18 (11%) patients who did not display a genuine PNH clone according to the international recommendations (at least 20 clustered events). For 7 of 163 selected patients, exchanges with centers are still ongoing for FCM data validation or due to missing clinical or biological data. Sixty-nine of the 138 validated patients displayed at least one follow-up point (Figure 1). Thus, we were able to compare 49 patients with rare PNH cells and 89 patients with a minor PNH clone with a significantly different clone size on the 3 lineages (median size on neutrophils: 0.03 [0.01 - 0.09] vs 0.34 [0.1 - 0.97] ; p<0.0001, on monocytes: 0.03 [0 - 7.88 ] vs 0.50 [0 - 7.08] ; p<0.0001, on red blood cells: 0.01 [0 - 1.36 ] vs 0.08 [0 - 1.24]; p= 0.0096). The patient median age at diagnosis was similar for both groups, 59 years [6 - 91] with 4 pediatric cases (<18y) for "rare PNH cells” group vs 49 years [10 - 87] with 9 pediatric cases for "minor PNH clone” group and a M/F sex ratio of 1.45 vs 0.98. Diagnoses were made between 2010 and 2021 with clinical information available in 98% of cases (n=47 and 88). Ninety-two to 94% of patients from both groups displayed BM failure including mainly aplastic anemia (n=30/47 vs 61/88), myelodysplastic syndrome (n=2 vs 6) and unexplained cytopenia(s) (n=11 vs 16). Only 4 cases of thrombosis for "rare PNH cells” group vs 3 for "minor PNH clone” group were observed, and the 2 cases of hemolytic anemia belong to the latter group. A more severe thrombopenia was observed in the "minor PNH clone” (68 [2-356] vs 44 [2-308] G/l) while all other biological parameters showed no significant difference. Long-term follow-up up to 7.0 years showed that one third of patients from both groups displayed an increase in clone size (above 0.1% for "rare PNH cells” and above 1% for "minor PNH clone", see Figure 2), most remained stable and a few decreased, the disappearance being always due to BM transplantation. Conclusion This multicenter study based on robust FCM analysis showed no differences between the clinical and biological presentation of patients with either rare PNH cells or a minor PNH clone. Long-term follow-up showed that rare PNH cells can evolve into minor PNH clone, suggesting that these cases may be considered as genuine minor clones. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Comparing Data Generated Using Multiparametric Flow Cytometry, Capillary Electrophoresis, and Next-Generation Sequencing in a Cohort of B-Cell Lymphoproliferative Disorder Samples

Background: Determining whether suspected lymphoid proliferation disorder (LPD) samples are benign or malignant remains a significant challenge for the hematopathologist. For the vast majority of LPDs, the current gold standard includes assessment of clinical characteristics, cytomorphology, and immunophenotyping by flow cytometry; however, 10-15% of LPDs require further resolution by molecular methods, such as PCR-based capillary electrophoresis or next-generation sequencing, to rule out clonality based on V-D-J rearrangements in the immunoglobulin heavy-chain (IGH) gene. Invivoscribe (San Diego, CA) has developed multiple complementary methodologies to evaluate LPDs; including multiparametric flow cytometry (MFC), capillary electrophoresis (CE), and next-generation sequencing (NGS). To compare performance between these three methods, a cohort of 21 clinical samples were tested in a pilot study. Method: The CAP/CLIA Hematolymphoid Screening assay, a 10-color flow panel, was used to identify all major types of hematopoietic neoplasia according to the 2006 Bethesda Consensus and 2017 WHO Classification. For molecular testing, DNA was extracted from peripheral blood (PB), then PCR-based assays were used to determine presence or absence of clonality. The Research Use Only (RUO) molecular assays were designed for CE (ABI 3500xL) and NGS (MiSeqTM) detection, utilizing framework multiplex PCR master mixes targeting the conserved framework regions FR1, FR2, or FR3 within the VH and the JH regions of the IGH gene. Twenty-one (21) de-identified PB samples were collected from informed consensus donors (LabPMM, San Diego, CA) or from residual samples (Lahey Hospital, Burlington, MA) and tested under an IRB approved protocol. An aliquot of each PB sample was used for MFC evaluation, then the remaining DNA was extracted and tested using the CE and NGS methods. Test results were reported as "present abnormal" or "normal" for MFC and "clonal" or "non-clonal" for molecular assays based on pre-defined criteria. Results: The MFC assay identified 12 out of 21 (57%) samples as "present abnormal" and 9 out of 21 (43%) samples as "normal." Both molecular assays (CE and NGS) reported 12 out of 21 (57%) samples as "clonal" and 9 out of 21 (43%) samples as "non-clonal." The results between CE and NGS assays were 100% in agreement for clonality detection. Comparisons between the CE and MFC assays again resulted in 100% agreement; all 12 "present abnormal" samples by MFC were reported as "clonal" by CE, while all 9 "normal" samples by MFC were "non-clonal" by CE. In addition to presence of clonality, the NGS method was able to detect the prevalence of a specific clonal V-D-J rearrangement sequence (expressed as % reads), which ranged from 10% to 99% in the 12 clonal samples. This pilot study demonstrated excellent concordance (21/21) between these 3 methods (MFC, CE and NGS). Accordingly, MFC, CE and NGS are all suitable methods for the evaluation and stratification of B-cell LPD samples. The combined use of all 3 methods provides a comprehensive assessment of clonality: including surface markers identified by MFC, DNA rearrangement lengths from CE, and the specific DNA rearrangement sequences from NGS. In addition, the NGS method allows for tracking of these clonal sequences in relation to disease status. Conclusion: This pilot study demonstrated 100% concordance between 3 different methods (MFC, CE, and NGS) when evaluating samples from LPD subjects. Depending on laboratory capabilities and the type of information desired, any one of these technologies can provide appropriate results when studying clonality. The alignment between MFC and NGS allows the combined use of both methods to provide rapid diagnosis via the MFC service and identification of specific biomarkers for subsequent tracking of these clonal populations in clinical studies.

Efficient transformer with code token learner for code clone detection

Deep learning techniques have achieved promising results in code clone detection in the past decade. Unfortunately, current deep learning-based methods rarely explicitly consider the modeling of long codes. Worse, the code length is increasing due to the increasing requirement of complex functions. Thus, modeling the relationship between code tokens to catch their long-range dependencies is crucial to comprehensively capture the information of the code fragment. In this work, we resort to the Transformer to capture long-range dependencies within a code, which however requires huge computational cost for long code fragments. To make it possible to apply Transformer efficiently, we propose a code token learner to largely reduce the number of feature tokens in an automatic way. Besides, considering the tree structure of the abstract syntax tree, we present a tree-based position embedding to encode the position of each token in the input. Apart from the Transformer that captures the dependency within a code, we further leverage a cross-code attention module to capture the similarities between two code fragments. Our method significantly reduces the computational cost of using Transformer by 97% while achieves superior performance with state-of-the-art methods. Our code is available at https://github.com/ArcticHare105/Code-Token-Learner.

A collaborative method for code clone detection using a deep learning model

Code cloning (CC) is the process of copying and reconfiguring a code fragment and using it in another part of a software project. This clones increases the running overhead of the software. As a result, Code Clone Detection (CCD) has become an active research area in software development research. The detection of Large-Variance Code Clones (LV-CCs) is very difficult when the lines of codes (LOCs) in the source code are very large. The distance metrics have been used in LV-CC detection by calculating the distance between training feature sets of source codes and testing feature sets of source codes. However, threshold selection for detecting clones is a challenging issue in distance-based LV-CC detection. To solve this, a Collaborative CCD using Deep Learning (CCCD-DL) is developed in this paper by utilising lexical, syntactic, semantic and structural features for identifying all types of clones together. A lexical feature is extracted from Clone Pairs (CPs) identified by LV-Mapper. Syntactic and semantic features are identified by the Abstract Syntax Tree (AST) and Control Flow Graph (CFG). The structural features are extracted by code size metrics (CZMs) and object-oriented metrics (OOMs). All features are coordinated and fed into the input layer of DNN. The hidden layer then transforms the inputs into the neural vertices in the multi-classification stage using linear transformation preceded by suppressing non-linearity. This process can generate a complicated and non-hypothetical prototype with a weight matrix for fitting the training sequence. Thus, the feed-forward step has been successfully completed. This model then uses back-propagation in the following element to modify the weight matrix based on the training set. Finally, a softmax layer converts the clone detection task into a classification process. The results of the experiments show that the proposed method solves distance-based problems more quickly and effectively than the traditional methods for the CCD.

Ranking Code Clones to Support Maintenance Activities

Ranking Code Clones to Support Maintenance Activities

Effects of Long Noncoding RNA HOXA-AS2 on the Proliferation and Migration of Gallbladder Cancer Cells.

To explore the function and mechanism of lncRNA HOXA-AS2 in cancer-associated fibroblasts (CAFs)-derived exosomes in gallbladder cancer metastasis, and provide new research targets for the treatment of gallbladder cancer. At the same time, in order to clarify the early predictive value of lncRNA HOXA-AS2 for gallbladder cancer metastasis, and to provide a theoretical basis for clinical individualized treatment of gallbladder cancer. Methods. In our previous work, we used TCGA database analysis to find that lncRNA HOXA-AS2 was highly expressed in gallbladder cancer tissues compared with normal tissues. In this study, the expression levels of HOXA-AS2 in gallbladder cancer cell lines and control cells were first verified by QPCR and Western blot methods. Then, lentiviral tools were used to construct knockdown vectors (RNAi#1, RNAi#2) and negative control vectors targeting two different sites of HOXA-AS2, and the vectors were transfected into NOZ and OCUG-1 cells, respectively. Real-time PCR was used to detect knockdown efficiency. Then, the effects of silencing HOXA-AS2 on the proliferation, cell viability, cell migration, and invasion ability of gallbladder cancer cells were detected by MTT, plate cloning assay, Transwell migration chamber assay, and Transwell invasion chamber assay. Finally, the interaction between HOXA-AS2 and miR-6867 and the 3′UTR of YAP1 protein was detected by luciferase reporter gene. The results showed that the expression level of HOXA-AS2 in gallbladder cancer cell lines was higher than that in control cells. The expression of HOXA-AS2 in gallbladder carcinoma tissues was significantly higher than that in adjacent tissues (p < 0.05). After successful knockout of HOXA-AS2 by lentiviral transfection, the expression of HOXA-AS2 in gallbladder cancer cell lines was significantly decreased. Through cell proliferation and plate clone detection, it was found that silencing HOXA-AS2 inhibited cell proliferation and invasion. Through software prediction and fluorescein reporter gene detection, it was found that HOXA-AS2 has a binding site with miR-6867, and the two are negatively correlated, that is, the expression of miR-6867 is enhanced after the expression of HOXA-AS2 is downregulated. And the 3′UTR of YAP1 protein in the Hippo signaling pathway binds to miR-6867. Therefore, HOXA-AS2 may affect the expression of YAP1 protein by regulating miR-6867, thereby inhibiting the Hippo signaling pathway and promoting the proliferation and metastasis of gallbladder cancer cells. HOXA-AS2 is abnormally expressed in gallbladder cancer cells. HOXA-AS2 may promote the migration and invasion of gallbladder cancer cells by regulating the Hippo signaling pathway through miR-6867. HOXA-AS2 may serve as a potential diagnostic and therapeutic target for gallbladder cancer in clinic.

Modeling Functional Similarity in Source Code With Graph-Based Siamese Networks

Code clones are duplicate code fragments that share (nearly) similar syntax or semantics. Code clone detection plays an important role in software maintenance, code refactoring, and reuse. A substantial amount of research has been conducted in the past to detect clones. A majority of these approaches use lexical and syntactic information to detect clones. However, only a few of them target semantic clones. Recently, motivated by the success of deep learning models in other fields, including natural language processing and computer vision, researchers have attempted to adopt deep learning techniques to detect code clones. These approaches use lexical information (tokens) and(or) syntactic structures like abstract syntax trees (ASTs) to detect code clones. However, they do not make sufficient use of the available structural and semantic information, hence limiting their capabilities. This paper addresses the problem of semantic code clone detection using program dependency graphs and geometric neural networks, leveraging the structured syntactic and semantic information. We have developed a prototype tool <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Holmes</small> , based on our novel approach and empirically evaluated it on popular code clone benchmarks. Our results show that <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Holmes</small> performs considerably better than the other state-of-the-art tool, TBCCD. We also assessed <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Holmes</small> on unseen projects and performed cross dataset experiments to evaluate the generalizability of <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Holmes</small> . Our results affirm that <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Holmes</small> outperforms TBCCD since most of the pairs that <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Holmes</small> detected were either undetected or suboptimally reported by TBCCD.

ASTENS-BWA: Searching partial syntactic similar regions between source code fragments via AST-based encoded sequence alignment

Code search is a common activity in software development, and code-to-code search can benefit in a wide range of use-case scenarios. Code-to-code search uses a code fragment as the query for searching similar code fragments from large corpora. The results of a search can be applied to some software engineering tasks, such as search-based code recommendation, data-driven program repairing, and software plagiarism detection. To be put into daily use, the code-to-code search needs to find similar code fragments accurately and efficiently in a large dataset. Some search engines can locate exactly similar code, but are not able to search syntactical clones. Therefore, we propose ASTENS-BWA, a novel approach for searching syntactic similar code regions between code fragments via a tree-based sequence alignment. Source code has been transformed into a tree-based sequence that contains the structure information, and a sequence alignment algorithm has been applied to find similar regions. We evaluate ASTENS-BWA on three different tasks, the results demonstrate that our approach can find syntactical similar regions for programming code and retrieve similar code fragments fast and with high accuracy. As a code clone detection tool, ASTENS-BWA can report clone pairs in a high recall, but it needs manually check to reduce the false alarms. ASTENS-BWA is scalable and can report cloned code fragments in seconds for a code corpus of million lines of code. • Novel code clone search approach that scales to large source code repositories. • Quick searching syntactical clones between code fragments via tree-based representation. • Code clone search approach that can outperform state-of-the-art techniques. • Application of the proposed approach towards code recommendation and clone detection.

Detection of Clones B2-ST131-C2 and A-ST617 in Escherichia coli Producing Both CTX-M-15 and CTX-M-27 from Tunisian Community Patients.

During a two-month period (2017–2018), 336 urine samples positive for Escherichia coli were collected from Tunisian patients. Of the 336 samples, 266 were collected from community patients and 70 from hospital settings. In all, 15 ESBL producers were identified (8 and 7, respectively) and assigned to 13 pulsotypes, including four ESBL-producing E. coli (ESBL-E) with E1 and E2 profiles (2 isolates each) from community patients. The two strains E1 were identified as B2-ST131 subclade C2 and the two isolates E2, A-ST617. The four strains carrying both CTX-M-15 and CTX-M-27, exhibited the multireplicon IncFII/F1A/F1B with the same formula F31:A4:B1. Two isolates with patterns E3 and E4 (Dice coefficient, 78.7%) isolated from community and hospital settings of two geographic areas were assigned to the emerging ST131 C1-M27 subclade and contained the replicon F1:A-:B20. The remaining ESBL-E divided into different sequence types/associated CTX-M: 2 ST131-C2/CTX-M-15 and ST744/CTX-M-55, ST617/CTM-15, ST2973/CTX-M-55, ST6448/CTX-M-15, ST224/CTX-M-15, ST1431/CTX-M-15, and ST38/CTX-M-27, one isolate each. Our study reports for the first time the presence in the Tunisian community of two clones of E. coli, including the virulent clone ST131-C2 harboring both CTX-M-15 and CTX-M-27, and confirms the spread of the emergent clone ST131-C1-M-27, notably in community urinary tract infections.

An interview with Miryung Kim - 2022 SIGSOFT Awardee

Miryung Kim received the 2022 SIGSOFT Influential Educator Award for outstanding achievements in undergraduate and graduate mentoring with emphasis on research excellence in software engineering, diversity, and inclusion. She received a Ph.D. in computer science from the University of Washington, Seattle (USA), and she is a Professor and a Vice Chair of graduate studies at UCLA Computer Science. Her research interests include the detection of code clones and code duplication, as well as solutions for their management and removal. She has helped to define the field of software engineering for data intensive computing.

Immunophenotypic assessment of PNH clones in major and minor cell lineages in the peripheral blood of patients with paroxysmal nocturnal hemoglobinuria.

Flow cytometric immunophenotyping is essential for the diagnosis of paroxysmal nocturnal hemoglobinuria (PNH). Most cases have easy to interpret flow cytometry profiles with red cells, neutrophils and monocytes showing complete deficiency of glycophosphatidylinositol (GPI) linked antigen expression. Some cases are more challenging to interpret due to the presence of multiple populations of PNH cells and variable levels of GPI antigen expression. We studied 46 known PNH patients, many with complex immunophenotypic profiles using a novel, single tube, multi-parameter 7-color immunophenotyping assay that allowed simultaneous detection and assessment of PNH clones within multiple lineages of peripheral blood leucocytes. Red cell PNH clones were also assessed in total and immature (CD71+) components by CD59 expression. For individual patients, total PNH clones in each cell lineage were highly correlated. Monocytes, eosinophils and basophils showed the highest proportions of PNH cells. Red cell PNH clones were typically smaller than monocyte and neutrophil PNH clones. In most cases, PNH clones were detectable in minor leucocyte populations where multiple populations of PNH cells were present, variability in the proportions of type II and type III cells was seen across different cell lineages, even though total PNH clones remained similar. This study shows that PNH patients with multiple PNH clones do not always display the same abnormality across all cell lineages routinely tested. There is no simple explanation for this but is likely due to a combination of complex molecular, genetic and biochemical dysfunction in different blood cell types.

Eksperimen Pengujian Optimizer dan Fungsi Aktivasi Pada Code Clone Detection dengan Pemanfaatan Deep Neural Network (DNN)

The problem of similarity (similarity) of program code can be one solution to the plagiarism detection approach. Plagiarism raises a form of action and consequences of plagiarism itself if the source used is not open source. Plagiarism is an act of deception of the work of others without the knowledge of the original author, which violates a Copyright and Moral Rights. With the increasing amount of data and data complexity, deep learning provides solutions for predictive analytics, with increased processing capabilities and optimal processor utilization. Deep learning shows success and improves the classification model in this field. On the other hand, clone detection code with massive, varied and high-speed data volumes requires feature extraction. With the potential of deep learning to extract better features, deep learning techniques are suitable for code clone detection. For this reason, it is necessary to develop a clone detection code that can process data from a programming language by utilizing deep learning. Based on the results of experiments conducted on 100 PHP program code data files, experimented with several types of activation function and optimizer methods. The average value of the resulting accuracy is good. With a variety of activation functions that we use such as Relu, Linear, Sigmoid, Softmax, Tanh, Elu, Selu, Softplus, Softsign, hard, and sigmoid, as well as variations of the optimizer used are Adagrad, RMSProp, SGD, Adadelta, Adam, Adamax and Nadam , the best attribute selection is in the Selu function and the RMSProp optimizer. The number of epochs used is 1000, the number of neurons per layer is 500 and the best number of hidden layers is 10, the average accuracy is 0.900