MOSAIC: An Artificial Intelligence-Based Framework for Multimodal Analysis, Classification, and Personalized Prognostic Assessment in Rare Cancers.

PURPOSE: Rare cancers constitute over 20% of human neoplasms, often affecting patients with unmet medical needs. The development of effective classification and prognostication systems is crucial to improve the decision-making process and drive innovative treatment strategies. We have created and implemented MOSAIC, an artificial intelligence (AI)-based framework designed for multimodal analysis, classification, and personalized prognostic assessment in rare cancers. Clinical validation was performed on myelodysplastic syndrome (MDS), a rare hematologic cancer with clinical and genomic heterogeneities. METHODS: We analyzed 4,427 patients with MDS divided into training and validation cohorts. Deep learning methods were applied to integrate and impute clinical/genomic features. Clustering was performed by combining Uniform Manifold Approximation and Projection for Dimension Reduction + Hierarchical Density-Based Spatial Clustering of Applications with Noise (UMAP + HDBSCAN) methods, compared with the conventional Hierarchical Dirichlet Process (HDP). Linear and AI-based nonlinear approaches were compared for survival prediction. Explainable AI (Shapley Additive Explanations approach [SHAP]) and federated learning were used to improve the interpretation and the performance of the clinical models, integrating them into distributed infrastructure. RESULTS: UMAP + HDBSCAN clustering obtained a more granular patient stratification, achieving a higher average silhouette coefficient (0.16) with respect to HDP (0.01) and higher balanced accuracy in cluster classification by Random Forest (92.7% ± 1.3% and 85.8% ± 0.8%). AI methods for survival prediction outperform conventional statistical techniques and the reference prognostic tool for MDS. Nonlinear Gradient Boosting Survival stands in the internal (Concordance-Index [C-Index], 0.77; SD, 0.01) and external validation (C-Index, 0.74; SD, 0.02). SHAP analysis revealed that similar features drove patients' subgroups and outcomes in both training and validation cohorts. Federated implementation improved the accuracy of developed models. CONCLUSION: MOSAIC provides an explainable and robust framework to optimize classification and prognostic assessment of rare cancers. AI-based approaches demonstrated superior accuracy in capturing genomic similarities and providing individual prognostic information compared with conventional statistical methods. Its federated implementation ensures broad clinical application, guaranteeing high performance and data protection.

N-of-one differential gene expression without control samples using a deep generative model.

Differential analysis of bulk RNA-seq data often suffers from lack of good controls. Here, we present a generative model that replaces controls, trained solely on healthy tissues. The unsupervised model learns a low-dimensional representation and can identify the closest normal representation for a given disease sample. This enables control-free, single-sample differential expression analysis. In breast cancer, we demonstrate how our approach selects marker genes and outperforms a state-of-the-art method. Furthermore, significant genes identified by the model are enriched in driver genes across cancers. Our results show that the in silico closest normal provides a more favorable comparison than control samples.

Synthetic Data Generation by Artificial Intelligence to Accelerate Research and Precision Medicine in Hematology.

PURPOSE: Synthetic data are artificial data generated without including any real patient information by an algorithm trained to learn the characteristics of a real source data set and became widely used to accelerate research in life sciences. We aimed to (1) apply generative artificial intelligence to build synthetic data in different hematologic neoplasms; (2) develop a synthetic validation framework to assess data fidelity and privacy preservability; and (3) test the capability of synthetic data to accelerate clinical/translational research in hematology. METHODS: A conditional generative adversarial network architecture was implemented to generate synthetic data. Use cases were myelodysplastic syndromes (MDS) and AML: 7,133 patients were included. A fully explainable validation framework was created to assess fidelity and privacy preservability of synthetic data. RESULTS: We generated MDS/AML synthetic cohorts (including information on clinical features, genomics, treatment, and outcomes) with high fidelity and privacy performances. This technology allowed resolution of lack/incomplete information and data augmentation. We then assessed the potential value of synthetic data on accelerating research in hematology. Starting from 944 patients with MDS available since 2014, we generated a 300% augmented synthetic cohort and anticipated the development of molecular classification and molecular scoring system obtained many years later from 2,043 to 2,957 real patients, respectively. Moreover, starting from 187 MDS treated with luspatercept into a clinical trial, we generated a synthetic cohort that recapitulated all the clinical end points of the study. Finally, we developed a website to enable clinicians generating high-quality synthetic data from an existing biobank of real patients. CONCLUSION: Synthetic data mimic real clinical-genomic features and outcomes, and anonymize patient information. The implementation of this technology allows to increase the scientific use and value of real data, thus accelerating precision medicine in hematology and the conduction of clinical trials.

Circular DNA in the human germline and its association with recombination.

Extrachromosomal circular DNA (eccDNA) is common in somatic tissue, but its existence and effects in the human germline are unexplored. We used microscopy, long-read DNA sequencing, and new analytic methods to document thousands of eccDNAs from human sperm. EccDNAs derived from all genomic regions and mostly contained a single DNA fragment, although some consisted of multiple fragments. The generation of eccDNA inversely correlates with the meiotic recombination rate, and chromosomes with high coding-gene density and Alu element abundance form the least eccDNA. Analysis of insertions in human genomes further indicates that eccDNA can persist in the human germline when the circular molecules reinsert themselves into the chromosomes. Our results suggest that eccDNA has transient and permanent effects on the germline. They explain how differences in the physical and genetic map might arise and offer an explanation of how Alu elements coevolved with genes to protect genome integrity against deleterious mutations producing eccDNA.

Replicative aging is associated with loss of genetic heterogeneity from extrachromosomal circular DNA in Saccharomyces cerevisiae.

Circular DNA can arise from all parts of eukaryotic chromosomes. In yeast, circular ribosomal DNA (rDNA) accumulates dramatically as cells age, however little is known about the accumulation of other chromosome-derived circles or the contribution of such circles to genetic variation in aged cells. We profiled circular DNA in Saccharomyces cerevisiae populations sampled when young and after extensive aging. Young cells possessed highly diverse circular DNA populations but 94% of the circular DNA were lost after ∼15 divisions, whereas rDNA circles underwent massive accumulation to >95% of circular DNA. Circles present in both young and old cells were characterized by replication origins including circles from unique regions of the genome and repetitive regions: rDNA and telomeric Y' regions. We further observed that circles can have flexible inheritance patterns: [HXT6/7circle] normally segregates to mother cells but in low glucose is present in up to 50% of cells, the majority of which must have inherited this circle from their mother. Interestingly, [HXT6/7circle] cells are eventually replaced by cells carrying stable chromosomal HXT6 HXT6/7 HXT7 amplifications, suggesting circular DNAs are intermediates in chromosomal amplifications. In conclusion, the heterogeneity of circular DNA offers flexibility in adaptation, but this heterogeneity is remarkably diminished with age.

Near-Random Distribution of Chromosome-Derived Circular DNA in the Condensed Genome of Pigeons and the Larger, More Repeat-Rich Human Genome.

Extrachromosomal circular DNA (eccDNA) elements of chromosomal origin are known to be common in a number of eukaryotic species. However, it remains to be addressed whether genomic features such as genome size, the load of repetitive elements within a genome, and/or animal physiology affect the number of eccDNAs. Here, we investigate the distribution and numbers of eccDNAs in a condensed and less repeat-rich genome compared with the human genome, using Columba livia domestica (domestic rock pigeon) as a model organism. By sequencing eccDNA in blood and breast muscle from three pigeon breeds at various ages and with different flight behavior, we characterize 30,000 unique eccDNAs. We identify genomic regions that are likely hotspots for DNA circularization in breast muscle, including genes involved in muscle development. We find that although eccDNA counts do not correlate with the biological age in pigeons, the number of unique eccDNAs in a nonflying breed (king pigeons) is significantly higher (9-fold) than homing pigeons. Furthermore, a comparison between eccDNA from skeletal muscle in pigeons and humans reveals ∼9-10 times more unique eccDNAs per human nucleus. The fraction of eccDNA sequences, derived from repetitive elements, exist in proportions to genome content, that is, human 72.4% (expected 52.5%) and pigeon 8.7% (expected 5.5%). Overall, our results support that eccDNAs are common in pigeons, that the amount of unique eccDNA types per nucleus can differ between species as well as subspecies, and suggest that eccDNAs from repeats are found in proportions relative to the content of repetitive elements in a genome.

Sensitive detection of circular DNAs at single-nucleotide resolution using guided realignment of partially aligned reads.

BACKGROUND: Circular DNA has recently been identified across different species including human normal and cancerous tissue, but short-read mappers are unable to align many of the reads crossing circle junctions hence limiting their detection from short-read sequencing data. RESULTS: Here, we propose a new method, Circle-Map that guides the realignment of partially aligned reads using information from discordantly mapped reads to map the short unaligned portions using a probabilistic model. We compared Circle-Map to similar up-to-date methods for circular DNA and RNA detection and we demonstrate how the approach implemented in Circle-Map dramatically increases sensitivity for detection of circular DNA on both simulated and real data while retaining high precision. CONCLUSION: Circle-Map is an easy-to-use command line tool that implements the required pipeline to accurately detect circular DNA from circle enriched next generation sequencing experiments. Circle-Map is implemented in python3.6 and it is freely available at https://github.com/iprada/Circle-Map.

To Be or Not to Be: Circular RNAs or mRNAs From Circular DNAs?

In recent years, there has been a growing interest in circular RNAs (circRNAs) since they are involved in a wide spectrum of cellular functions that might have a large impact on phenotype and disease. CircRNAs are mainly recorded by RNA-Seq and computational methods focused on the detection of back-splicing junction sequences considered the diagnostic feature of circRNAs. While some protocols remove linear RNA prior to sequencing, many have characterized circRNAs by sorting through total RNA sequencing data without excluding the possibility that some linear RNA can provide the same signal as a circRNA. Recent studies have revealed that circular DNAs of chromosomal origin are common in eukaryotic genomes and that they can be transcribed. Transcription events across the junction of circular DNAs would result in a transcript with a junction similar to those present in circRNAs. Therefore, in this report, we want to draw attention to transcripts from such circular DNAs both as an interesting new player in the transcriptome and also as a confounding factor that must be taken into account when studying circRNAs.

Circular DNA elements of chromosomal origin are common in healthy human somatic tissue.

The human genome is generally organized into stable chromosomes, and only tumor cells are known to accumulate kilobase (kb)-sized extrachromosomal circular DNA elements (eccDNAs). However, it must be expected that kb eccDNAs exist in normal cells as a result of mutations. Here, we purify and sequence eccDNAs from muscle and blood samples from 16 healthy men, detecting ~100,000 unique eccDNA types from 16 million nuclei. Half of these structures carry genes or gene fragments and the majority are smaller than 25 kb. Transcription from eccDNAs suggests that eccDNAs reside in nuclei and recurrence of certain eccDNAs in several individuals implies DNA circularization hotspots. Gene-rich chromosomes contribute to more eccDNAs per megabase and the most transcribed protein-coding gene in muscle, TTN (titin), provides the most eccDNAs per gene. Thus, somatic genomes are rich in chromosome-derived eccDNAs that may influence phenotypes through altered gene copy numbers and transcription of full-length or truncated genes.

Circular RNA profiling reveals that circular RNAs from ANXA2 can be used as new biomarkers for multiple sclerosis.

Multiple sclerosis is an autoimmune disease, with higher prevalence in women, in whom the immune system is dysregulated. This dysregulation has been shown to correlate with changes in transcriptome expression as well as in gene-expression regulators, such as non-coding RNAs (e.g. microRNAs). Indeed, some of these have been suggested as biomarkers for multiple sclerosis even though few biomarkers have reached the clinical practice. Recently, a novel family of non-coding RNAs, circular RNAs, has emerged as a new player in the complex network of gene-expression regulation. MicroRNA regulation function through a 'sponge system' and a RNA splicing regulation function have been proposed for the circular RNAs. This regulating role together with their high stability in biofluids makes them seemingly good candidates as biomarkers. Given the dysregulation of both protein-coding and non-coding transcriptome that have been reported in multiple sclerosis patients, we hypothesised that circular RNA expression may also be altered. Therefore, we carried out expression profiling of 13.617 circular RNAs in peripheral blood leucocytes from multiple sclerosis patients and healthy controls finding 406 differentially expressed (P-value < 0.05, Fold change > 1.5) and demonstrate after validation that, circ_0005402 and circ_0035560 are underexpressed in multiple sclerosis patients and could be used as biomarkers of the disease.