EpiDiP/NanoDiP: a versatile unsupervised machine learning edge computing platform for epigenomic tumour diagnostics.

DNA methylation analysis based on supervised machine learning algorithms with static reference data, allowing diagnostic tumour typing with unprecedented precision, has quickly become a new standard of care. Whereas genome-wide diagnostic methylation profiling is mostly performed on microarrays, an increasing number of institutions additionally employ nanopore sequencing as a faster alternative. In addition, methylation-specific parallel sequencing can generate methylation and genomic copy number data. Given these diverse approaches to methylation profiling, to date, there is no single tool that allows (1) classification and interpretation of microarray, nanopore and parallel sequencing data, (2) direct control of nanopore sequencers, and (3) the integration of microarray-based methylation reference data. Furthermore, no software capable of entirely running in routine diagnostic laboratory environments lacking high-performance computing and network infrastructure exists. To overcome these shortcomings, we present EpiDiP/NanoDiP as an open-source DNA methylation and copy number profiling suite, which has been benchmarked against an established supervised machine learning approach using in-house routine diagnostics data obtained between 2019 and 2021. Running locally on portable, cost- and energy-saving system-on-chip as well as gpGPU-augmented edge computing devices, NanoDiP works in offline mode, ensuring data privacy. It does not require the rigid training data annotation of supervised approaches. Furthermore, NanoDiP is the core of our public, free-of-charge EpiDiP web service which enables comparative methylation data analysis against an extensive reference data collection. We envision this versatile platform as a useful resource not only for neuropathologists and surgical pathologists but also for the tumour epigenetics research community. In daily diagnostic routine, analysis of native, unfixed biopsies by NanoDiP delivers molecular tumour classification in an intraoperative time frame.

Clinical, histopathological and molecular features of dedifferentiated melanomas: An EORTC Melanoma Group Retrospective Analysis.

PURPOSE: Dedifferentiated melanoma (DedM) poses significant diagnostic challenges. We aimed to investigate the clinical, histopathological and molecular features of DedM. Methylation signature (MS) and copy number profiling (CNP) were carried out in a subgroup of cases. PATIENTS AND METHODS: A retrospective series of 78 DedM tissue samples from 61 patients retrieved from EORTC (European Organisation for Research and Treatment of Cancer) Melanoma Group centres were centrally reviewed. Clinical and histopathological features were retrieved. In a subgroup of patients, genotyping through Infinium Methylation microarray and CNP analysis was carried out. RESULTS: Most patients (60/61) had a metastatic DedM showing most frequently an unclassified pleomorphic, spindle cell, or small round cell morphology akin to undifferentiated soft tissue sarcoma, rarely associated with heterologous elements. Overall, among 20 successfully analysed tissue samples from 16 patients, we found retained melanoma-like MS in only 7 tissue samples while a non-melanoma-like MS was observed in 13 tissue samples. In two patients from whom multiple specimens were analysed, some of the samples had a preserved cutaneous melanoma MS while other specimens exhibited an epigenetic shift towards a mesenchymal/sarcoma-like profile, matching the histological features. In these two patients, CNP was largely identical across all analysed specimens, in line with their common clonal origin, despite significant modification of their epigenome. CONCLUSIONS: Our study further highlights that DedM represents a real diagnostic challenge. While MS and genomic CNP may help pathologists to diagnose DedM, we provide proof-of-concept that dedifferentiation in melanoma is frequently associated with epigenetic modifications.

Alterations of pre-mRNA splicing in human inflammatory bowel disease.

Alternative pre-mRNA splicing is regarded as a pivotal mechanism for generating proteome diversity and complexity from a limited inventory of mammalian genes. Aberrant splicing has been described as a predisposing factor for a number of diseases, but very little is known about its role in chronic inflammation. In this study, we systematically screened 149 splicing factors and 145 potential intron retention events for occurrence and differential expression in inflammatory bowel diseases (IBD). As a result, we identified 47 splicing factors and 33 intron retention events that were differentially regulated in mucosal tissue of IBD patients at transcript level. Despite the fact that Crohn's disease and ulcerative colitis, two subtypes of IBD, share the expression patterns of splicing factors and intron retention events in the majority of cases, we observed significant differences. To investigate these subtype-specific changes in detail we determined the expression levels of seven splicing factors (DUSP11, HNRPAB, HNRPH3, SLU7, SFR2IP, SFPQ, SF3B14) and three intron retention events (PARC, IER3, FGD2) in a cohort of 165 patients with inflammatory diseases of the colon (120 with IBD) and 30 healthy controls by real time PCR (TaqMan). This study demonstrates the potential impact of regulated splicing factors on subsequent regulated intron retention in the pathogenesis of chronic inflammation, exemplified by IBD.

Positioning of necrotic lobular intraepithelial neoplasias (LIN, grade 3) within the sequence of breast carcinoma progression.

Lobular intraepithelial neoplasia Grade 3 (LIN3) is a recently recognized variant of intraepithelial lobular neoplasia (LIN) of the breast composed of either uniform, generally small cells with massive lobular distension, pleomorphic cells, signet-ring cells, or any cell type with necrosis. In contrast to classic forms of LIN, there is no consensus on therapeutic strategies for LIN3. In part this is due to the paucity of molecular data that could assist in defining the relationship of LIN3 to classic LIN and carcinomas. In this study we have employed array comparative genomic hybridization to determine the patterns of chromosomal aberrations in nine LIN3 lesions. By comparison to array CGH data of 13 classic LIN lesions, we demonstrate that classic LIN and LIN3 share several recurrent changes, in particular gains of 1q and losses of 16q. Both aberrations are known to appear early in tumorigenesis and to be associated with good prognosis. However, apart from this overlap, there were a number of karyotypic features that were observed exclusively in LIN3. Clearly, this lesion was characterized by a significantly higher number of DNA copy number changes (9 vs. 31 on average), a considerable complexity of chromosomal rearrangements with more than 16 breakpoints in one chromosome and overlapping high copy amplifications encompassing a number of known oncogenes. Our data suggest that, at the genetic level, LIN3 represents a highly advanced lesion with considerable resemblance to carcinomas and, therefore, might represent the transition state from an intraepithelial neoplasm to breast carcinoma.

Age-specific hormonal decline is accompanied by transcriptional changes in human sebocytes in vitro.

The importance of hormones in endogenous aging has been displayed by recent studies performed on animal models and humans. To decipher the molecular mechanisms involved in aging we maintained human sebocytes at defined hormone-substituted conditions that corresponded to average serum levels of females from 20 (f20) to 60 (f60) years of age. The corresponding hormone receptor expression was demonstrated by reverse transcription-polymerase chain reaction (RT-PCR), Western blotting and immunocytochemistry. Cells at f60 produced significantly lower lipids than at f20. Increased mRNA and protein levels of c-Myc and increased protein levels of FN1, which have been associated with aging, were detected in SZ95 sebocytes at f60 compared to those detected at f20 after 5 days of treatment. Expression profiling employing a cDNA microarray composed of 15 529 cDNAs identified 899 genes with altered expression levels at f20 vs. f60. Confirmation of gene regulation was performed by real-time RT-PCR. The functional annotation of these genes according to the Gene Ontology identified pathways related to mitochondrial function, oxidative stress, ubiquitin-mediated proteolysis, cell cycle, immune responses, steroid biosynthesis and phospholipid degradation - all hallmarks of aging. Twenty-five genes in common with those identified in aging kidneys and several genes involved in neurodegenerative diseases were also detected. This is the first report describing the transcriptome of human sebocytes and its modification by a cocktail of hormones administered in age-specific levels and provides an in vitro model system, which approximates some of the hormone-dependent changes in gene transcription that occur during aging in humans.

High throughput identification of potential Arabidopsis mitogen-activated protein kinases substrates.

Mitogen-activated protein kinase (MAPK) cascades are universal and highly conserved signal transduction modules in eucaryotes, including plants. These protein phosphorylation cascades link extracellular stimuli to a wide range of cellular responses. However, the underlying mechanisms are so far unknown as information about phosphorylation substrates of plant MAPKs is lacking. In this study we addressed the challenging task of identifying potential substrates for Arabidopsis thaliana mitogen-activated protein kinases MPK3 and MPK6, which are activated by many environmental stress factors. For this purpose, we developed a novel protein microarray-based proteomic method allowing high throughput study of protein phosphorylation. We generated protein microarrays including 1,690 Arabidopsis proteins, which were obtained from the expression of an almost nonredundant uniclone set derived from an inflorescence meristem cDNA expression library. Microarrays were incubated with MAPKs in the presence of radioactive ATP. Using a threshold-based quantification method to evaluate the microarray results, we were able to identify 48 potential substrates of MPK3 and 39 of MPK6. 26 of them are common for both kinases. One of the identified MPK6 substrates, 1-aminocyclopropane-1-carboxylic acid synthase-6, was just recently shown as the first plant MAPK substrate in vivo, demonstrating the potential of our method to identify substrates with physiological relevance. Furthermore we revealed transcription factors, transcription regulators, splicing factors, receptors, histones, and others as candidate substrates indicating that regulation in response to MAPK signaling is very complex and not restricted to the transcriptional level. Nearly all of the 48 potential MPK3 substrates were confirmed by other in vitro methods. As a whole, our approach makes it possible to shortlist candidate substrates of mitogen-activated protein kinases as well as those of other protein kinases for further analysis. Follow-up in vivo experiments are essential to evaluate their physiological relevance.

Systematic delineation of a calmodulin peptide interaction.

We present a comprehensive profile of amino acid side-chain constraints in a calmodulin (CaM) peptide complex. These data were obtained from the analysis of calmodulin binding to an array of all single substitution analogues as well as N- and C-terminal truncations of the skMLCK derived M13 peptide ligand. The experimentally derived binding data were evaluated with respect to the known 3D-structure of the CaM/M13 complex. Besides an almost perfect agreement between the measured affinities and the structural data, the unexpected high-affine Asn5Ala variant of the M13(*) peptide described by Montigiani et al. could be verified. In contrast to other reports our data clearly support the postulate of the minor and major hydrophobic anchors of this calcium dependent interaction.

Multiplexed sorting of libraries on libraries: a novel method for empirical protein design by affinity-driven phage enrichment on synthetic peptide arrays.

Chemically synthesized peptide arrays on planar cellulose carriers are proposed as libraries of ligands suitable for the multiplexed simultaneous capture of peptide-specific acceptor proteins from a large randomly mutagenized library of acceptor proteins presented on bacteriophage M13 particles. This experimental set-up can be exploited to rapidly screen for individual new, distinct binding partners from two complementary libraries (two-dimensional screening). The technical feasibility of this empirical protein design approach was demonstrated with calmodulin as an aceptor protein using an array of mastoparan variants for multiplexed phage affinity enrichment.