An optimized protocol for quality control of gene therapy vectors using nanopore direct RNA sequencing.

Despite recent advances made toward improving the efficacy of lentiviral gene therapies, a sizeable proportion of produced vector contains an incomplete and thus potentially nonfunctional RNA genome. This can undermine gene delivery by the lentivirus as well as increase manufacturing costs and must be improved to facilitate the widespread clinical implementation of lentiviral gene therapies. Here, we compare three long-read sequencing technologies for their ability to detect issues in vector design and determine nanopore direct RNA sequencing to be the most powerful. We show how this approach identifies and quantifies incomplete RNA caused by cryptic splicing and polyadenylation sites, including a potential cryptic polyadenylation site in the widely used Woodchuck Hepatitis Virus Posttranscriptional Regulatory Element (WPRE). Using artificial polyadenylation of the lentiviral RNA, we also identify multiple hairpin-associated truncations in the analyzed lentiviral vectors (LVs), which account for most of the detected RNA fragments. Finally, we show that these insights can be used for the optimization of LV design. In summary, nanopore direct RNA sequencing is a powerful tool for the quality control and optimization of LVs, which may help to improve lentivirus manufacturing and thus the development of higher quality lentiviral gene therapies.

CD8+ tissue-resident memory T cells are expanded in primary Sjögren's disease and can be therapeutically targeted by CD103 blockade.

OBJECTIVE: Tissue-resident memory cells (Trm) are a subset of T cells residing persistently and long-term within specific tissues that contribute to persistent inflammation and tissue damage. We characterised the phenotype and function of Trm and the role of CD103 in primary Sjogren's syndrome (pSS). METHODS: In both pSS and non-pSS sicca syndrome patients, we examined Trm frequency, cytokine production in salivary glands (SG) and peripheral blood (PB). We also analysed Trm-related gene expression in SG biopsies through bulk and single-cell RNA sequencing (scRNAseq). Additionally, we investigated Trm properties in an immunisation-induced animal model of pSS (experimental SS, ESS) mouse model and assessed the effects of Trm inhibition via intraglandular anti-CD103 monoclonal antibody administration. RESULTS: Transcriptomic pSS SG showed an upregulation of genes associated with tissue recruitment and long-term survival of Trm cells, confirmed by a higher frequency of CD8+CD103+CD69+ cells in pSS SG, compared with non-specific sialadenitis (nSS). In SG, CD8+ CD103+ Trm contributed to the secretion of granzyme-B and interferon-γ, CD8+ Trm cells were localised within inflammatory infiltrates, where PD1+CD8+ T cells were also increased compared with nSS and MALT lymphoma. scRNAseq of PB and pSS SG T cells confirmed expression of CD69, ITGAE, GZMB, GZMK and HLA-DRB1 among CD3+CD8+ SG T cells. In the SG of ESS, CD8+CD69+CD103+ Trm producing Granzyme B progressively expanded. However, intraglandular blockade of CD103 in ESS reduced Trm, reduced glandular damage and improved salivary flow. CONCLUSIONS: CD103+CD8+Trm cells are expanded in the SG of pSS and ESS, participate in tissue inflammation and can be therapeutically targeted.

Management of patients with advanced epithelial ovarian cancer: a European survey.

BACKGROUND: The aim of this study was to assess current European practices in the management of patients with advanced epithelial ovarian cancer in 2021. METHODS: A 58-question electronic survey was distributed anonymously to the members of six European learned societies. Initial diagnostic workup and staging, pathological data, surgical data, treatments and follow-up strategies were assessed. RESULTS: A total of 171 participants from 17 European countries responded to emailed surveys. Most participants were experienced practitioners (superior than 15 years of experience) specializing in gynecology-obstetrics (29.8%), surgical oncology (25.1%), and oncogynecology (21.6%). According to most (64.8%) participants, less than 50% of patients were eligible for primary debulking surgery. Variations in the rate of primary debulking surgery depending on the country of origin of the practitioners were observed in this study. The LION study criteria were applied in 70.4% of cases during PDS and 27.1% after chemotherapy. In cases of BRCA1-2 mutations, olaparib was given by 75.0-84.8% of respondents, whereas niraparib was given in cases of BRCA wild-type diseases. CONCLUSIONS: This study sheds light on current practices and attitudes regarding the management of patients with advanced epithelial ovarian cancer in Europe in 2021.

Unique Transcriptional Architecture in Airway Epithelial Cells and Macrophages Shapes Distinct Responses following Influenza Virus Infection Ex Vivo.

Airway epithelial cells and macrophages differ markedly in their responses to influenza A virus (IAV) infection. To investigate transcriptional responses underlying these differences, purified subsets of type II airway epithelial cells (ATII) and alveolar macrophages (AM) recovered from the lungs of mock- or IAV-infected mice at 9 h postinfection were subjected to RNA sequencing. This time point was chosen to allow for characterization of cell types first infected with the virus inoculum, prior to multicycle virus replication and the infiltration of inflammatory cells into the airways. In the absence of infection, AM predominantly expressed genes related to immunity, whereas ATII expressed genes consistent with their physiological roles in the lung. Following IAV infection, AM almost exclusively activated cell-intrinsic antiviral pathways that were dependent on interferon (IFN) regulatory factor 3/7 (IRF3/7) and/or type I IFN signaling. In contrast, IAV-infected ATII activated a broader range of physiological responses, including cell-intrinsic antiviral pathways, which were both independent of and dependent on IRF3/7 and/or type I IFN. These data suggest that transcriptional profiles hardwired during development are a major determinant underlying the different responses of ATII and AM to IAV infection.IMPORTANCE Airway epithelial cells (AEC) and airway macrophages (AM) represent major targets of influenza A virus (IAV) infection in the lung, yet the two cell types respond very differently to IAV infection. We have used RNA sequencing to define the host transcriptional responses in each cell type under steady-state conditions as well as following IAV infection. To do this, different cell subsets isolated from the lungs of mock- and IAV-infected mice were subjected to RNA sequencing. Under steady-state conditions, AM and AEC express distinct transcriptional activities, consistent with distinct physiological roles in the airways. Not surprisingly, these cells also exhibited major differences in transcriptional responses following IAV infection. These studies shed light on how the different transcriptional architectures of airway cells from two different lineages drive transcriptional responses to IAV infection.

Exploring the feasibility and utility of exome-scale tumour sequencing in a clinical setting.

BACKGROUND: Technology has progressed from single gene panel to large-scale genomic sequencing. This is raising expectations from clinicians and patients alike. The utility and performance of this technology in a clinical setting needs to be evaluated. AIM: This pilot study investigated the feasibility of using exome-scale sequencing (ESS) to identify molecular drivers within cancers in real-time for Precision Oncology in the clinic. METHODS: Between March 2014 and March 2015, the Victorian Comprehensive Cancer Centre Alliance explored the feasibility and utility of ESS in a pilot study. DNA extracted from the tumour specimens underwent both ESS and targeted 'hotspot' sequencing (TS). Blood was taken for germline analysis. A multi-disciplinary molecular tumour board determined the clinical relevance of identified mutations; in particular, whether they were 'actionable' and/or 'druggable'. RESULTS: Of 23 patients screened, 15 (65%) met the tissue requirements for genomic analysis. TS and ESS were successful in all cases. ESS identified pathogenic somatic variants in 73% (11/15 cases) versus 53% (8/15 cases) using TS. Clinically focused ESS identified 63 variants, consisting of 30 somatic variants (including all 13 identified by TS) and 33 germline variants. Overall, there were 48 unique variants. ESS had a clinical impact in 53% (8/15 cases); 47% (7/15 cases) were referred to the familial cancer clinic, and 'druggable' targets were identified in 53% (8/15 cases). CONCLUSION: ESS of tumour DNA impacted clinical decision-making in 53%, with 20% more pathogenic variants identified through ESS than TS. The identification of germline variants in 47% was an unexpected finding.

Circulating tumour cells from patients with colorectal cancer have cancer stem cell hallmarks in ex vivo culture.

OBJECTIVE: Although counting of circulating tumour cells (CTC) has attracted a broad interest as potential markers of tumour progression and treatment response, the lack of functional characterisation of these cells had become a bottleneck in taking these observations to the clinic. Our objective was to culture these cells in order to understand them and exploit their therapeutic potential to the full. DESIGN: Here, hypothesising that some CTC potentially have cancer stem cell (CSC) phenotype, we generated several CTC lines from the blood of patients with advanced metastatic colorectal cancer (CRC) based on their self-renewal abilities. Multiple standard tests were then employed to characterise these cells. RESULTS: Our CTC lines self-renew, express CSC markers and have multilineage differentiation ability, both in vitro and in vivo. Patient-derived CTC lines are tumorigenic in subcutaneous xenografts and are also able to colonise the liver after intrasplenic injection. RNA sequencing analyses strikingly demonstrate that drug metabolising pathways represent the most upregulated feature among CTC lines in comparison with primary CRC cells grown under similar conditions. This result is corroborated by the high resistance of the CTC lines to conventional cytotoxic compounds. CONCLUSIONS: Taken together, our results directly demonstrate the existence of patient-derived colorectal CTCs that bear all the functional attributes of CSCs. The CTC culture model described here is simple and takes <1 month from blood collection to drug testing, therefore, routine clinical application could facilitate access to personalised medicine. CLINICAL TRIAL REGISTRATION: ClinicalTrial.gov NCT01577511.

AmpliVar: mutation detection in high-throughput sequence from amplicon-based libraries.

Conventional means of identifying variants in high-throughput sequencing align each read against a reference sequence, and then call variants at each position. Here, we demonstrate an orthogonal means of identifying sequence variation by grouping the reads as amplicons prior to any alignment. We used AmpliVar to make key-value hashes of sequence reads and group reads as individual amplicons using a table of flanking sequences. Low-abundance reads were removed according to a selectable threshold, and reads above this threshold were aligned as groups, rather than as individual reads, permitting the use of sensitive alignment tools. We show that this approach is more sensitive, more specific, and more computationally efficient than comparable methods for the analysis of amplicon-based high-throughput sequencing data. The method can be extended to enable alignment-free confirmation of variants seen in hybridization capture target-enrichment data.

Socrates: identification of genomic rearrangements in tumour genomes by re-aligning soft clipped reads.

MOTIVATION: Methods for detecting somatic genome rearrangements in tumours using next-generation sequencing are vital in cancer genomics. Available algorithms use one or more sources of evidence, such as read depth, paired-end reads or split reads to predict structural variants. However, the problem remains challenging due to the significant computational burden and high false-positive or false-negative rates. RESULTS: In this article, we present Socrates (SOft Clip re-alignment To idEntify Structural variants), a highly efficient and effective method for detecting genomic rearrangements in tumours that uses only split-read data. Socrates has single-nucleotide resolution, identifies micro-homologies and untemplated sequence at break points, has high sensitivity and high specificity and takes advantage of parallelism for efficient use of resources. We demonstrate using simulated and real data that Socrates performs well compared with a number of existing structural variant detection tools. AVAILABILITY AND IMPLEMENTATION: Socrates is released as open source and available from http://bioinf.wehi.edu.au/socrates CONTACT: papenfuss@wehi.edu.au Supplementary information: Supplementary data are available at Bioinformatics online.

Transcriptomic analysis supports similar functional roles for the two thymuses of the tammar wallaby.

BACKGROUND: The thymus plays a critical role in the development and maturation of T-cells. Humans have a single thoracic thymus and presence of a second thymus is considered an anomaly. However, many vertebrates have multiple thymuses. The tammar wallaby has two thymuses: a thoracic thymus (typically found in all mammals) and a dominant cervical thymus. Researchers have known about the presence of the two wallaby thymuses since the 1800s, but no genome-wide research has been carried out into possible functional differences between the two thymic tissues. Here, we used pyrosequencing to compare the transcriptomes of a cervical and thoracic thymus from a single 178 day old tammar wallaby. RESULTS: We show that both the tammar thoracic and the cervical thymuses displayed gene expression profiles consistent with roles in T-cell development. Both thymuses expressed genes that mediate distinct phases of T-cells differentiation, including the initial commitment of blood stem cells to the T-lineage, the generation of T-cell receptor diversity and development of thymic epithelial cells. Crucial immune genes, such as chemokines were also present. Comparable patterns of expression of non-coding RNAs were seen. 67 genes differentially expressed between the two thymuses were detected, and the possible significance of these results are discussed. CONCLUSION: This is the first study comparing the transcriptomes of two thymuses from a single individual. Our finding supports that both thymuses are functionally equivalent and drive T-cell development. These results are an important first step in the understanding of the genetic processes that govern marsupial immunity, and also allow us to begin to trace the evolution of the mammalian immune system.

Sensitive molecular testing methods can demonstrate NSCLC driver mutations in malignant pleural effusion despite non-malignant cytology.

Malignant pleural effusion (MPE) may be diagnosed by cytologic evaluation of pleural fluid, though false negative results can occur. Pleural effusions may provide a source of tumour material for genotyping in lung cancer patients. Detection of MPE may be improved through use of highly sensitive molecular techniques. We identified five patients with non-small cell lung cancer (NSCLC) with initial pleural fluid samples that were non-malignant on cytology, but were subsequently clinically confirmed to have MPE. Tumour mutation status was confirmed via routine testing of diagnostic clinical specimens. Cytologically negative pleural fluid cell-block specimens were analysed by amplicon-based parallel sequencing (APS) for somatic mutations commonly detected in NSCLC, and selected cases by improved and complete enrichment CO-amplification at lower denaturation temperature PCR (ICECOLD PCR) for known mutations. Mutations were detected in three out of three (sensitivity 100%) cytologically non-malignant pleural fluids from patients with a known mutation: two patients with known Kirsten rat sarcoma (KRAS) mutation demonstrated the same KRAS mutation in their pleural fluids by APS, both at approximately 2% mutant allele frequency. In one patient with a known KRAS mutation, ICECOLD PCR detected the same KRAS variant at 0.7% frequency. No mutations were detected in patients with wild-type findings from reference samples (specificity 100%). Sensitive DNA sequencing methods can detect cancer-driver mutations in cytologically non-malignant pleural fluid specimens from NSCLC patients with MPE. Our findings demonstrate the feasibility of sensitive molecular diagnostic techniques for improvement of diagnostic assessment of pleural effusions in patients with lung cancer.

Clinical Utility of Real-Time Targeted Molecular Profiling in the Clinical Management of Ovarian Cancer: The ALLOCATE Study.

PURPOSE: The ALLOCATE study was designed as a pilot to demonstrate the feasibility and clinical utility of real-time targeted molecular profiling of patients with recurrent or advanced ovarian cancer for identification of potential targeted therapies. PATIENTS AND METHODS: A total of 113 patients with ovarian cancer of varying histologies were recruited from two tertiary hospitals, with 99 patient cases suitable for prospective analysis. Targeted molecular and methylation profiling of fresh biopsy and archived tumor samples were performed by screening for mutations or copy-number variations in 44 genes and for promoter methylation of BRCA1 and RAD51C. RESULTS: Somatic genomic or methylation events were identified in 85% of all patient cases, with potentially actionable events with defined targeted therapies (including four resistance events) detected in 60% of all patient cases. On the basis of these findings, six patients received molecularly guided therapy, three patients had unsuspected germline cancer-associated BRCA1/2 mutations and were referred for genetic counseling, and two intermediate differentiated (grade 2) serous ovarian carcinomas were reclassified as low grade, leading to changes in clinical management. Additionally, secondary reversion mutations in BRCA1/2 were identified in fresh biopsy samples of two patients, consistent with clinical platinum/poly (ADP-ribose) polymerase inhibitor resistance. Timely reporting of results if molecular testing is done at disease recurrence, as well as early referral for patients with platinum-resistant cancers, were identified as factors that could improve the clinical utility of molecular profiling. CONCLUSION: ALLOCATE molecular profiling identified known genomic and methylation alterations of the different ovarian cancer subtypes and was deemed feasible and useful in routine clinical practice. Better patient selection and access to a wider range of targeted therapies or clinical trials will further enhance the clinical utility of molecular profiling.

Binning sequences using very sparse labels within a metagenome.

BACKGROUND: In metagenomic studies, a process called binning is necessary to assign contigs that belong to multiple species to their respective phylogenetic groups. Most of the current methods of binning, such as BLAST, k-mer and PhyloPythia, involve assigning sequence fragments by comparing sequence similarity or sequence composition with already-sequenced genomes that are still far from comprehensive. We propose a semi-supervised seeding method for binning that does not depend on knowledge of completed genomes. Instead, it extracts the flanking sequences of highly conserved 16S rRNA from the metagenome and uses them as seeds (labels) to assign other reads based on their compositional similarity. RESULTS: The proposed seeding method is implemented on an unsupervised Growing Self-Organising Map (GSOM), and called Seeded GSOM (S-GSOM). We compared it with four well-known semi-supervised learning methods in a preliminary test, separating random-length prokaryotic sequence fragments sampled from the NCBI genome database. We identified the flanking sequences of the highly conserved 16S rRNA as suitable seeds that could be used to group the sequence fragments according to their species. S-GSOM showed superior performance compared to the semi-supervised methods tested. Additionally, S-GSOM may also be used to visually identify some species that do not have seeds. The proposed method was then applied to simulated metagenomic datasets using two different confidence threshold settings and compared with PhyloPythia, k-mer and BLAST. At the reference taxonomic level Order, S-GSOM outperformed all k-mer and BLAST results and showed comparable results with PhyloPythia for each of the corresponding confidence settings, where S-GSOM performed better than PhyloPythia in the >/= 10 reads datasets and comparable in the > or = 8 kb benchmark tests. CONCLUSION: In the task of binning using semi-supervised learning methods, results indicate S-GSOM to be the best of the methods tested. Most importantly, the proposed method does not require knowledge from known genomes and uses only very few labels (one per species is sufficient in most cases), which are extracted from the metagenome itself. These advantages make it a very attractive binning method. S-GSOM outperformed the binning methods that depend on already-sequenced genomes, and compares well to the current most advanced binning method, PhyloPythia.

Using growing self-organising maps to improve the binning process in environmental whole-genome shotgun sequencing.

Metagenomic projects using whole-genome shotgun (WGS) sequencing produces many unassembled DNA sequences and small contigs. The step of clustering these sequences, based on biological and molecular features, is called binning. A reported strategy for binning that combines oligonucleotide frequency and self-organising maps (SOM) shows high potential. We improve this strategy by identifying suitable training features, implementing a better clustering algorithm, and defining quantitative measures for assessing results. We investigated the suitability of each of di-, tri-, tetra-, and pentanucleotide frequencies. The results show that dinucleotide frequency is not a sufficiently strong signature for binning 10 kb long DNA sequences, compared to the other three. Furthermore, we observed that increased order of oligonucleotide frequency may deteriorate the assignment result in some cases, which indicates the possible existence of optimal species-specific oligonucleotide frequency. We replaced SOM with growing self-organising map (GSOM) where comparable results are obtained while gaining 7%-15% speed improvement.

Evaluating the Genomic Yield of a Single Endobronchial Ultrasound-guided Transbronchial Needle Aspiration in Lung Cancer: Meeting the Challenge of Doing More With Less.

BACKGROUND: Minimally invasive techniques, including endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA), yield small specimens that are adequate for cytologic diagnosis of lung cancer, but also need to provide material for molecular analysis to guide treatment. The number of EBUS-TBNA passes needed for mutation testing remains unclear. We sought to assess the adequacy of a single pass for genomic profiling of actionable mutations. METHODS: In a prospective observational study, paired samples from the same lesion were obtained from patients undergoing EBUS-TBNA for lung cancer diagnosis/staging. Following tumor cell confirmation by rapid on-site evaluation, a "reference" sample comprising ≥ 3 passes was obtained and formalin-fixed paraffin-embedded. A "study" sample comprising a single pass was taken and snap-frozen. The primary outcome was DNA yield and quality from a single pass. The secondary outcome was diagnostic accuracy of a single pass for detecting actionable mutations. RESULTS: In 40 patients, single-pass specimens yielded a mean 3.98 µg of highly intact DNA, well above the minimum threshold for targeted sequencing, which was performed in adenocarcinoma cases (n = 24). In 23 cases, there was 100% agreement in mutation status between reference and study samples. In 1 case, the reference sample failed to generate a molecular diagnosis owing to insufficient tumor cells; however, the study specimen identified a KRAS mutation. Tumor cell percentage in mutation-positive specimens was 1% to 70%, suggesting that single-pass samples detect mutations even when tumor cell content is low. CONCLUSION: Single EBUS-TBNA passes yield DNA of high quantity and quality with high accuracy for molecular profiling, irrespective of tumor cell content.

High-Throughput Amplicon-Based Copy Number Detection of 11 Genes in Formalin-Fixed Paraffin-Embedded Ovarian Tumour Samples by MLPA-Seq.

Whilst next generation sequencing can report point mutations in fixed tissue tumour samples reliably, the accurate determination of copy number is more challenging. The conventional Multiplex Ligation-dependent Probe Amplification (MLPA) assay is an effective tool for measurement of gene dosage, but is restricted to around 50 targets due to size resolution of the MLPA probes. By switching from a size-resolved format, to a sequence-resolved format we developed a scalable, high-throughput, quantitative assay. MLPA-seq is capable of detecting deletions, duplications, and amplifications in as little as 5ng of genomic DNA, including from formalin-fixed paraffin-embedded (FFPE) tumour samples. We show that this method can detect BRCA1, BRCA2, ERBB2 and CCNE1 copy number changes in DNA extracted from snap-frozen and FFPE tumour tissue, with 100% sensitivity and >99.5% specificity.

The architecture and evolution of cancer neochromosomes.

We isolated and analyzed, at single-nucleotide resolution, cancer-associated neochromosomes from well- and/or dedifferentiated liposarcomas. Neochromosomes, which can exceed 600 Mb in size, initially arise as circular structures following chromothripsis involving chromosome 12. The core of the neochromosome is amplified, rearranged, and corroded through hundreds of breakage-fusion-bridge cycles. Under selective pressure, amplified oncogenes are overexpressed, while coamplified passenger genes may be silenced epigenetically. New material may be captured during punctuated chromothriptic events. Centromeric corrosion leads to crisis, which is resolved through neocentromere formation or native centromere capture. Finally, amplification terminates, and the neochromosome core is stabilized in linear form by telomere capture. This study investigates the dynamic mutational processes underlying the life history of a special form of cancer mutation.

Novel venom gene discovery in the platypus.

BACKGROUND: To date, few peptides in the complex mixture of platypus venom have been identified and sequenced, in part due to the limited amounts of platypus venom available to study. We have constructed and sequenced a cDNA library from an active platypus venom gland to identify the remaining components. RESULTS: We identified 83 novel putative platypus venom genes from 13 toxin families, which are homologous to known toxins from a wide range of vertebrates (fish, reptiles, insectivores) and invertebrates (spiders, sea anemones, starfish). A number of these are expressed in tissues other than the venom gland, and at least three of these families (those with homology to toxins from distant invertebrates) may play non-toxin roles. Thus, further functional testing is required to confirm venom activity. However, the presence of similar putative toxins in such widely divergent species provides further evidence for the hypothesis that there are certain protein families that are selected preferentially during evolution to become venom peptides. We have also used homology with known proteins to speculate on the contributions of each venom component to the symptoms of platypus envenomation. CONCLUSIONS: This study represents a step towards fully characterizing the first mammal venom transcriptome. We have found similarities between putative platypus toxins and those of a number of unrelated species, providing insight into the evolution of mammalian venom.

The Tasmanian devil transcriptome reveals Schwann cell origins of a clonally transmissible cancer.

The Tasmanian devil, a marsupial carnivore, is endangered because of the emergence of a transmissible cancer known as devil facial tumor disease (DFTD). This fatal cancer is clonally derived and is an allograft transmitted between devils by biting. We performed a large-scale genetic analysis of DFTD with microsatellite genotyping, a mitochondrial genome analysis, and deep sequencing of the DFTD transcriptome and microRNAs. These studies confirm that DFTD is a monophyletic clonally transmissible tumor and suggest that the disease is of Schwann cell origin. On the basis of these results, we have generated a diagnostic marker for DFTD and identify a suite of genes relevant to DFTD pathology and transmission. We provide a genomic data set for the Tasmanian devil that is applicable to cancer diagnosis, disease evolution, and conservation biology.

An unsupervised hierarchical dynamic self-organizing approach to cancer class discovery and marker gene identification in microarray data.

MOTIVATION: Current Self-Organizing Maps (SOMs) approaches to gene expression pattern clustering require the user to predefine the number of clusters likely to be expected. Hierarchical clustering methods used in this area do not provide unique partitioning of data. We describe an unsupervised dynamic hierarchical self-organizing approach, which suggests an appropriate number of clusters, to perform class discovery and marker gene identification in microarray data. In the process of class discovery, the proposed algorithm identifies corresponding sets of predictor genes that best distinguish one class from other classes. The approach integrates merits of hierarchical clustering with robustness against noise known from self-organizing approaches. RESULTS: The proposed algorithm applied to DNA microarray data sets of two types of cancers has demonstrated its ability to produce the most suitable number of clusters. Further, the corresponding marker genes identified through the unsupervised algorithm also have a strong biological relationship to the specific cancer class. The algorithm tested on leukemia microarray data, which contains three leukemia types, was able to determine three major and one minor cluster. Prediction models built for the four clusters indicate that the prediction strength for the smaller cluster is generally low, therefore labelled as uncertain cluster. Further analysis shows that the uncertain cluster can be subdivided further, and the subdivisions are related to two of the original clusters. Another test performed using colon cancer microarray data has automatically derived two clusters, which is consistent with the number of classes in data (cancerous and normal). AVAILABILITY: JAVA software of dynamic SOM tree algorithm is available upon request for academic use. SUPPLEMENTARY INFORMATION: A comparison of rectangular and hexagonal topologies for GSOM is available from http://www.mame.mu.oz.au/mechatronics/journalinfo/Hsu2003supp.pdf