Sequencing Depth Research Articles

Predicting environmental stressor levels with machine learning: a comparison between amplicon sequencing, metagenomics, and total RNA sequencing based on taxonomically assigned data

IntroductionMicrobes are increasingly (re)considered for environmental assessments because they are powerful indicators for the health of ecosystems. The complexity of microbial communities necessitates powerful novel tools to derive conclusions for environmental decision-makers, and machine learning is a promising option in that context. While amplicon sequencing is typically applied to assess microbial communities, metagenomics and total RNA sequencing (herein summarized as omics-based methods) can provide a more holistic picture of microbial biodiversity at sufficient sequencing depths. Despite this advantage, amplicon sequencing and omics-based methods have not yet been compared for taxonomy-based environmental assessments with machine learning.MethodsIn this study, we applied 16S and ITS-2 sequencing, metagenomics, and total RNA sequencing to samples from a stream mesocosm experiment that investigated the impacts of two aquatic stressors, insecticide and increased fine sediment deposition, on stream biodiversity. We processed the data using similarity clustering and denoising (only applicable to amplicon sequencing) as well as multiple taxonomic levels, data types, feature selection, and machine learning algorithms and evaluated the stressor prediction performance of each generated model for a total of 1,536 evaluated combinations of taxonomic datasets and data-processing methods.ResultsSequencing and data-processing methods had a substantial impact on stressor prediction. While omics-based methods detected a higher diversity of taxa than amplicon sequencing, 16S sequencing outperformed all other sequencing methods in terms of stressor prediction based on the Matthews Correlation Coefficient. However, even the highest observed performance for 16S sequencing was still only moderate. Omics-based methods performed poorly overall, but this was likely due to insufficient sequencing depth. Data types had no impact on performance while feature selection significantly improved performance for omics-based methods but not for amplicon sequencing.DiscussionWe conclude that amplicon sequencing might be a better candidate for machine-learning-based environmental stressor prediction than omics-based methods, but the latter require further research at higher sequencing depths to confirm this conclusion. More sampling could improve stressor prediction performance, and while this was not possible in the context of our study, thousands of sampling sites are monitored for routine environmental assessments, providing an ideal framework to further refine the approach for possible implementation in environmental diagnostics.

Improved detection of aberrant splicing with FRASER 2.0 and the intron Jaccard index

Detection of aberrantly spliced genes is an important step in RNA-seq-based rare-disease diagnostics. We recently developed FRASER, a denoising autoencoder-based method that outperformed alternative methods of detecting aberrant splicing. However, because FRASER's three splice metrics are partially redundant and tend to be sensitive to sequencing depth, we introduce here a more robust intron-excision metric, the intron Jaccard index, that combines the alternative donor, alternative acceptor, and intron-retention signal into a single value. Moreover, we optimized model parameters and filter cutoffs by using candidate rare-splice-disrupting variants as independent evidence. On 16,213 GTEx samples, our improved algorithm, FRASER 2.0, called typically 10 times fewer splicing outliers while increasing the proportion of candidate rare-splice-disrupting variants by 10-fold and substantially decreasing the effect of sequencing depth on the number of reported outliers. To lower the multiple-testing correction burden, we introduce an option to select the genes to be tested for each sample instead of a transcriptome-wide approach. This option can be particularly useful when prior information, such as candidate variants or genes, is available. Application on 303 rare-disease samples confirmed the relative reduction in the number of outlier calls for a slight loss of sensitivity; FRASER 2.0 recovered 22 out of 26 previously identified pathogenic splicing cases with default cutoffs and 24 when multiple-testing correction was limited to OMIM genes containing rare variants. Altogether, these methodological improvements contribute to more effective RNA-seq-based rare diagnostics by drastically reducing the amount of splicing outlier calls per sample at minimal loss of sensitivity.

Evaluating the performance of low-frequency variant calling tools for the detection of variants from short-read deep sequencing data

Detection of low-frequency variants with high accuracy plays an important role in biomedical research and clinical practice. However, it is challenging to do so with next-generation sequencing (NGS) approaches due to the high error rates of NGS. To accurately distinguish low-level true variants from these errors, many statistical variants calling tools for calling low-frequency variants have been proposed, but a systematic performance comparison of these tools has not yet been performed. Here, we evaluated four raw-reads-based variant callers (SiNVICT, outLyzer, Pisces, and LoFreq) and four UMI-based variant callers (DeepSNVMiner, MAGERI, smCounter2, and UMI-VarCal) considering their capability to call single nucleotide variants (SNVs) with allelic frequency as low as 0.025% in deep sequencing data. We analyzed a total of 54 simulated data with various sequencing depths and variant allele frequencies (VAFs), two reference data, and Horizon Tru-Q sample data. The results showed that the UMI-based callers, except smCounter2, outperformed the raw-reads-based callers regarding detection limit. Sequencing depth had almost no effect on the UMI-based callers but significantly influenced on the raw-reads-based callers. Regardless of the sequencing depth, MAGERI showed the fastest analysis, while smCounter2 consistently took the longest to finish the variant calling process. Overall, DeepSNVMiner and UMI-VarCal performed the best with considerably good sensitivity and precision of 88%, 100%, and 84%, 100%, respectively. In conclusion, the UMI-based callers, except smCounter2, outperformed the raw-reads-based callers in terms of sensitivity and precision. We recommend using DeepSNVMiner and UMI-VarCal for low-frequency variant detection. The results provide important information regarding future directions for reliable low-frequency variant detection and algorithm development, which is critical in genetics-based medical research and clinical applications.

Computational Pipeline for the Detection of Plant RNA Viruses Using High-Throughput Sequencing.

In this chapter, we describe a computational pipeline for the in silico detection of plant viruses by high-throughput sequencing (HTS) from total RNA samples. The pipeline is designed for the analysis of short reads generated using an Illumina platform and free-available software tools. First, we provide advice for high-quality total RNA purification, library preparation, and sequencing. The bioinformatics pipeline begins with the raw reads obtained from the sequencing machine and performs some curation steps to obtain long contigs. Contigs are blasted against a local database of reference nucleotide viral sequences to identify the viruses in the samples. Then, the search is refined by applying specific filters. We also provide the code to re-map the short reads against the viruses found to get information on sequencing depth and read coverage for each virus. No previous bioinformatics background is required, but basic knowledge of the Unix command line and R language is recommended.

Improved quality metrics for association and reproducibility in chromatin accessibility data using mutual information

BackgroundCorrelation metrics are widely utilized in genomics analysis and often implemented with little regard to assumptions of normality, homoscedasticity, and independence of values. This is especially true when comparing values between replicated sequencing experiments that probe chromatin accessibility, such as assays for transposase-accessible chromatin via sequencing (ATAC-seq). Such data can possess several regions across the human genome with little to no sequencing depth and are thus non-normal with a large portion of zero values. Despite distributed use in the epigenomics field, few studies have evaluated and benchmarked how correlation and association statistics behave across ATAC-seq experiments with known differences or the effects of removing specific outliers from the data. Here, we developed a computational simulation of ATAC-seq data to elucidate the behavior of correlation statistics and to compare their accuracy under set conditions of reproducibility.ResultsUsing these simulations, we monitored the behavior of several correlation statistics, including the Pearson’s R and Spearman’s rho coefficients as well as Kendall’s tau and Top–Down correlation. We also test the behavior of association measures, including the coefficient of determination R^2, Kendall’s W, and normalized mutual information. Our experiments reveal an insensitivity of most statistics, including Spearman’s rho, Kendall’s tau, and Kendall’s W, to increasing differences between simulated ATAC-seq replicates. The removal of co-zeros (regions lacking mapped sequenced reads) between simulated experiments greatly improves the estimates of correlation and association. After removing co-zeros, the R^2 coefficient and normalized mutual information display the best performance, having a closer one-to-one relationship with the known portion of shared, enhanced loci between simulated replicates. When comparing values between experimental ATAC-seq data using a random forest model, mutual information best predicts ATAC-seq replicate relationships.ConclusionsCollectively, this study demonstrates how measures of correlation and association can behave in epigenomics experiments. We provide improved strategies for quantifying relationships in these increasingly prevalent and important chromatin accessibility assays.

Modelling the effectiveness of surveillance based on metagenomics in detecting, monitoring, and forecasting antimicrobial resistance in livestock production under economic constraints

Current surveillance of antimicrobial resistance (AMR) is mostly based on testing indicator bacteria using minimum inhibitory concentration (MIC) panels. Metagenomics has the potential to identify all known antimicrobial resistant genes (ARGs) in complex samples and thereby detect changes in the occurrence earlier. Here, we simulate the results of an AMR surveillance program based on metagenomics in the Danish pig population. We modelled both an increase in the occurrence of ARGs and an introduction of a new ARG in a few farms and the subsequent spread to the entire population. To make the simulation realistic, the total cost of the surveillance was constrained, and the sampling schedule was set at one pool per month with 5, 20, 50, or 100 samples. Our simulations demonstrate that a pool of 20–50 samples and a sequencing depth of 250 million fragments resulted in the shortest time to detection in both scenarios, with a time delay to detection of change of >hspace{0.17em}15 months in all scenarios. Compared with culture-based surveillance, our simulation indicates that there are neither significant reductions nor increases in time to detect a change using metagenomics. The benefit of metagenomics is that it is possible to monitor all known resistance in one sampling and laboratory procedure in contrast to the current monitoring that is based on the phenotypic characterisation of selected indicator bacterial species. Therefore, overall changes in AMR in a population will be detected earlier using metagenomics due to the fact that the resistance gene does not have to be transferred to and expressed by an indicator bacteria before it is possible to detect.

Whole-genome sequencing and comparative genomics reveal the potential pathogenic mechanism of Neoscytalidium dimidiatum on pitaya

ABSTRACT Neoscytalidium dimidiatum (class Dothideomycetes ) is a fungus responsible for canker disease in pitaya stems and fruits, leading to significant economic losses. However, little is known about the pathogenesis, family evolution, and genetic variants of this species. In this study, we report a high-quality genome sequence of N. dimidiatum based on the Nanopore sequencing technology platform for sequencing and Hi-C assembly technology for genome assembly. The genome contains 12 chromosomes (2 n = 2× = 12; diploid), with a sequencing depth of 186.1×, encoding 12,349 proteins. Molecular phylogenetic analysis showed that N. dimidiatum is evolutionarily close to Botryosphaeria dothidea . Compared to other fungi, the N. dimidiatum genome contains many carbohydrate-active enzymes and secondary metabolites. Additionally, we predict that N. dimidiatum contains 121 candidate effectors that may play important roles in infection and colonization, promoting pathogenicity in pitaya. Nine of these effectors were confirmed to contain signal peptides and inhibit BAX/INF1-induced necrosis in Nicotiana benthamiana , demonstrating their importance during infection. Finally, we also confirm that N. dimidiatum does not form an appressorium or infection thread but instead infects pitaya via open stomata. In conclusion, the results provide a foundation for future research on N. dimidiatum and the control of pitaya canker. IMPORTANCE Pitaya canker is a significant disease in the pitaya industry in China, causing significant economic losses. Therefore, systematic research on Neoscytalidium dimidiatum , the fungus implicated in pitaya canker, is essential for comprehending the pathogenesis of this disease and developing effective control strategies. We applied comparative genomics to reveal the genetic evolution, metabolic diversity, environmental adaptation, and pathogenicity of N. dimidiatum , providing ideal targets for studies of pathogenesis and molecular targets for fungicide development. Moreover, the systematic study of the N. dimidiatum growth cycle, morphological characteristics, and molecular phylogenetic analysis can promote a comprehensive understanding of its genetic basis.

An efficient large‐scale whole‐genome sequencing analyses practice with an average daily analysis of 100Tbp: ZBOLT

AbstractBackgroundWith the advancement of whole‐genome sequencing (WGS) technology, massively parallel sequencing (MPS) remains the mainstream due to its accuracy, low cost, and high throughput. The development of the analytical pipeline corresponding to MPS has always been of great importance. Increasingly large population genomics studies, as a specific type of big data research, pose new challenges for analysis solutions.ResultsHere, we introduce ZBOLT, a comprehensive analysis system that incorporates both software and hardware advancements, making it an appropriate choice for large‐scale population genomic studies that require extensive data processing. In this study, we first evaluate ZBOLT's calling accuracy using the Genome in a Bottle (GIAB) benchmark dataset. Then we apply ZBOLT to a large‐scale population genomics study with 5,616 high sequencing depth samples totaling 1.16Pbp (base pair). As the results show, ZBOLT demonstrates exceptional efficiency and low energy consumption, processing 100Tbp per day and using 1kWh per 100Gbp sequenced sample.ConclusionThis research serves as a valuable reference for analyzing sequencing data from large population cohorts and underscores the significant potential of ZBOLT in large‐scale population genomics studies.

Neonatal Gut Microbiome Composition and Gut-Derived Bloodstream Infections in a Cohort of Premature Singaporean Infants

Abstract Background Due to their gestational age, low birth weight, and immature gut barriers, premature neonates are at high risk of bloodstream infections (BSI) caused by bacteria that reside in the gastrointestinal tract. Relatively little is known about the effect of alterations to the gut microbiota on the risk of BSI in premature neonates. Methods This prospective cohort study was conducted in the neonatal intensive care unit in the KK Women’s and Children’s Hospital in Singapore between June 1, 2019 and May 31, 2021. Clinical data and twice weekly fecal samples were collected during the first 75 days of life from infants born at or below 30 weeks gestational age. All infants received a Bifidobacterium breve probiotic starting at median (interquartile range [IQR]) day of life (DOL) 2 (1, 3). Fecal samples underwent shotgun metagenomic sequencing, and the resulting reads were aligned to Kraken 2 for taxonomic classification. We utilized permutational multivariate analysis of variance (PERMANOVA) and mixed effects linear regression to evaluate associations between clinical factors and gut microbiota composition. We described gut microbiota alterations among a subset of infants who developed a BSI. Results We collected 581 fecal samples from 75 neonates. Median (IQR) gestational age was 27 (25, 29) weeks and birthweight was 955 (763, 1210) grams. There were 18 BSIs that occurred among these infants and were caused by Streptococcus agalactiae (n=5), Klebsiella pneumoniae (n=3), Escherichia coli (n=2), Staphylococcus aureus (n=2), Staphylococcus epidermidis (n=2), Proteus mirabilis (n=1), Acinetobacter baumannii (n=1), Streptococcus anginosus (n=1), and Streptococcus pasteurianus (n=1). Metagenomic sequencing generated a median (IQR) sequencing depth of 5.9M (4.4M, 34.9M) paired-end reads per sample. The clinical factors that accounted for the most variance in gut microbiota composition were DOL (PERMANOVA; R2=0.022, p=0.001), feeding modality (R2=0.020, p=0.001), and antibiotic exposure (R2=0.018, p=0.001). Probiotic exposure was associated with an increase in the probiotic species, B. breve, and also a decrease in the relative abundances of several Clostridia species. Broad spectrum antibiotic exposures were associated with lower relative abundances of both potential pathogens, e.g., bacteria from the genera Enterobacter, Klebsiella, and Enterococcus, and beneficial bacteria, e.g., Bifidobacteria species, including the probiotic species. Notably, metronidazole and carbapenem exposures were associated with an increase in Staphylococcus species. Of 18 neonates who developed BSI, the causative species was detected in the gut microbiota of 17 infants. The relative abundance of the BSI species was ≥50% immediately prior to or at the time of BSI in 9 (50%) infants (Figure). Interestingly, four (22%) infants had a continued high abundance of the BSI species in the gut microbiota despite definitive antibiotic therapy. Conclusion Several clinical factors contribute to the gut microbiota composition of premature Singaporean neonates, especially postnatal age, feeding modality, and antibiotic exposures. Broad-spectrum antibiotic exposures affected the relative abundance of the probiotic agent and were associated with increases in staphylococcal species known to be potentially pathogenic in neonates. If detected in the gut microbiota, the bacteria that caused the BSI was often present in significant relative abundances both before and after the BSI. Figure. Relative abundances of bacterial species in the gut microbiota before and after bloodstream infections (BSI) in a subset of premature neonates. The relative abundances of the BSI species are plotted in color, whereas relative abundances of other detected species are depicted in gray. Each darkened vertical bar corresponds to a single fecal sample. Shading between each bar is provided to visualize changes in relative abundance between fecal samples. Black vertical lines depict the day of BSI. Specific antibiotic exposures are shown in the colored horizontal bars atop each subject’s time-series plot.

Accuracy and depth evaluation of clinical low pass genome sequencing in the detection of mosaic aneuploidies and CNVs

BackgroundLow-pass genome sequencing (LP GS) has shown distinct advantages over traditional methods for the detection of mosaicism. However, no study has systematically evaluated the accuracy of LP GS in the detection of mosaic aneuploidies and copy number variants (CNVs) in prenatal diagnosis. Moreover, the influence of sequencing depth on mosaicism detection of LP GS has not been fully evaluated.MethodsTo evaluate the accuracy of LP GS in the detection of mosaic aneuploidies and mosaic CNVs, 27 samples with known aneuploidies and CNVs and 1 negative female sample were used to generate 6 simulated samples and 21 virtual samples, each sample contained 9 different mosaic levels. Mosaic levels were simulated by pooling reads or DNA from each positive sample and the negative sample according to a series of percentages (ranging from 3 to 40%). Then, the influence of sequencing depth on LP GS in the detection of mosaic aneuploidies and CNVs was evaluated by downsampling.ResultsTo evaluate the accuracy of LP GS in the detection of mosaic aneuploidies and CNVs, a comparative analysis of mosaic levels was performed using 6 simulated samples and 21 virtual samples with 35 M million (M) uniquely aligned high-quality reads (UAHRs). For mosaic levels > 30%, the average difference (detected mosaic levels vs. theoretical mosaic levels) of 6 mosaic CNVs in simulated samples was 4.0%, and the average difference (detected mosaic levels vs. mosaic levels of Y chromosome) of 6 mosaic aneuploidies and 15 mosaic CNVs in virtual samples was 2.7%. Furthermore, LP GS had a higher detection rate and accuracy for the detection of mosaic aneuploidies and CNVs of larger sizes, especially mosaic aneuploidies. For depth evaluation, the results of LP GS in downsampling samples were compared with those of LP GS using 35 M UAHRs. The detection sensitivity of LP GS for 6 mosaic aneuploidies and 15 mosaic CNVs in virtual samples increased with UAHR. For mosaic levels > 30%, the total detection sensitivity reached a plateau at 30 M UAHRs. With 30 M UAHRs, the total detection sensitivity was 99.2% for virtual samples.ConclusionsWe demonstrated the accuracy of LP GS in mosaicism detection using simulated data and virtual samples, respectively. Thirty M UAHRs (single-end 35 bp) were optimal for LP GS in the detection of mosaic aneuploidies and most mosaic CNVs larger than 1.48 Mb (Megabases) with mosaic levels > 30%. These results could provide a reference for laboratories that perform clinical LP GS in the detection of mosaic aneuploidies and CNVs.

Normalizing need not be the norm: count-based math for analyzing single-cell data.

Counting transcripts of mRNA are a key method of observation in modern biology. With advances in counting transcripts in single cells (single-cell RNA sequencing or scRNA-seq), these data are routinely used to identify cells by their transcriptional profile, and to identify genes with differential cellular expression. Because the total number of transcripts counted per cell can vary for technical reasons, the first step of many commonly used scRNA-seq workflows is to normalize by sequencing depth, transforming counts into proportional abundances. The primary objective of this step is to reshape the data such that cells with similar biological proportions of transcripts end up with similar transformed measurements. But there is growing concern that normalization and other transformations result in unintended distortions that hinder both analyses and the interpretation of results. This has led to an intense focus on optimizing methods for normalization and transformation of scRNA-seq data. Here, we take an alternative approach, by avoiding normalization and transformation altogether. We abandon the use of distances to compare cells, and instead use a restricted algebra, motivated by measurement theory and abstract algebra, that preserves the count nature of the data. We demonstrate that this restricted algebra is sufficient to draw meaningful and practical comparisons of gene expression through the use of the dot product and other elementary operations. This approach sidesteps many of the problems with common transformations, and has the added benefit of being simpler and more intuitive. We implement our approach in the package countland, available in python and R.

INNV-36. ULTRA-FAST DEEP-LEARNED CNS TUMOR CLASSIFICATION USING NANOPORE-SEQUENCING DURING SURGERY

Abstract BACKGROUND Preoperative imaging and intraoperative histological analysis provide important guidance during CNS tumor surgery but are not always precise enough. Also, the diagnosis of CNS tumors increasingly requires a combination of morphological and molecular analysis. Illumina DNA methylation array analysis provides information on hundreds of thousands of methylation sites genome-wide but is relatively time-consuming and expensive. Nanopore sequencing enables real-time generation of an increasingly precise methylation profile. AIM: To exploit the potential of nanopore sequencing as a tool for intraoperative molecular diagnostics of CNS tumors. METHODS We developed Sturgeon, a patient-agnostic, transfer-learned neural network approach. Sturgeon leverages publicly available Illumina 450K DNA methylation profiles to simulate millions of nanopore sequencing experiments, which were used to train and validate classifier models. Subsequently, snap-frozen archival samples of 50 CNS tumors were analyzed, as well as fresh tumor samples of an additional 25 CNS tumors intraoperatively. RESULTS Sturgeon delivers models capable of classifying 81 different CNS tumor types based on relatively sparse, randomly distributed nanopore sequencing data. Sturgeon was able to suggest within 40 minutes after the start of sequencing the correct diagnosis with high confidence in 45 out of 50 archival CNS tumor samples; in the remaining cases Sturgeon abstained from a diagnosis. In the real-life scenario, in 18 out of 25 cases (72%) the suggested diagnosis was correct, and in 7 cases the required confidence threshold was not reached (mostly because of low tumor cell percentage). Importantly, the Sturgeon models are portable, work across patients and sequence depths, and require less than a minute of computation on a laptop CPU. CONCLUSIONS Intraoperative nanopore sequencing can assist neurosurgical decision making and opens avenues for starting tumor-specific therapy already during operation. Furthermore, this diagnostic approach has the potential to reduce the costs of methylation profiling analysis and speed-up the multidisciplinary decision-making process.

Key pathways and genes that are altered during treatment with hyperbaric oxygen in patients with sepsis due to necrotizing soft tissue infection (HBOmic study)

BackgroundFor decades, the basic treatment strategies of necrotizing soft tissue infections (NSTI) have remained unchanged, primarily relying on aggressive surgical removal of infected tissue, broad-spectrum antibiotics, and supportive intensive care. One treatment strategy that has been proposed as an adjunctive measure to improve patient outcomes is hyperbaric oxygen (HBO2) treatment. HBO2 treatment has been linked to several immune modulatory effects; however, investigating these effects is complicated due to the disease's acute life-threatening nature, metabolic and cell homeostasis dependent variability in treatment effects, and heterogeneity with respect to both patient characteristics and involved pathogens. To embrace this complexity, we aimed to explore the underlying biological mechanisms of HBO2 treatment in patients with NSTI on the gene expression level.MethodsWe conducted an observational cohort study on prospective collected data, including 85 patients admitted to the intensive care unit (ICU) for NSTI. All patients were treated with one or two HBO2 treatments and had one blood sample taken before and after the intervention. Total RNAs from blood samples were extracted and mRNA purified with rRNA depletion, followed by whole-transcriptome RNA sequencing with a targeted sequencing depth of 20 million reads. A model for differentially expressed genes (DEGs) was fitted, and the functional aspects of the obtained set of genes was predicted with GO (Gene Ontology) and KEGG (Kyoto Encyclopedia of genes and Genomes) enrichment analyses. All analyses were corrected for multiple testing with FDR.ResultsAfter sequential steps of quality control, a final of 160 biological replicates were included in the present study. We found 394 protein coding genes that were significantly DEGs between the two conditions with FDR < 0.01, of which 205 were upregulated and 189 were downregulated. The enrichment analysis of these DEGs revealed 20 GO terms in biological processes and 12 KEGG pathways that were significantly overrepresented in the upregulated DEGs, of which the term; “adaptive immune response” (GO:0002250) (FDR = 9.88E-13) and “T cell receptor signaling pathway” (hsa04660) (FDR = 1.20E-07) were the most significant. Among the downregulated DEGs two biological processes were significantly enriched, of which the GO term “apoptotic process” (GO:0006915) was the most significant (FDR = 0.001), followed by “Positive regulation of T helper 1 cell cytokine production” (GO:2000556), and “NF-kappa B signaling pathway” (hsa04064) was the only KEGG pathway that was significantly overrepresented (FDR = 0.001).ConclusionsWhen one or two sessions of HBO2 treatment were administered to patients with a dysregulated immune response and systemic inflammation due to NSTI, the important genes that were regulated during the intervention were involved in activation of T helper cells and downregulation of the disease-induced highly inflammatory pathway NF-κB, which was associated with a decrease in the mRNA level of pro-inflammatory factors.Trial registration: Biological material was collected during the INFECT study, registered at ClinicalTrials.gov (NCT01790698).

257 Effects of Prebiotic Galactooligosaccharide (GOS) on Hindgut Microbial Composition of Pigs During the Peri-Weaning Period

Abstract Galactooligosaccharide (GOS) is a prebiotic containing beta-linked galactose oligomers of 2 to 8 units. Previously, GOS positively impacted post-weaning growth performance and altered jejunal morphology. In this study, GOS-enriched whey permeate (Milk Specialties Global, Eden Prairie MN) was supplemented to piglets in farrowing and nursery phases. To maximize pre-weaning GOS intake, novel gruel creep feeders were used. Piglets from 24 litters received either no creep feed (NC), creep without GOS (CG-) or creep with 5.0% GOS (CG+) followed by a phase 1 nursery diet without (NG-) or with 3.8% GOS (NG+). At d 22 (pre-weaning) and d 31 (post-weaning), 6 pigs per treatment were euthanized for sample collection. DNA was extracted from cecal and rectal swabs for downstream 16S rRNA sequencing via Illumina MiSeq (Zymo Research Corp, Irvine CA). An amplicon sequence variant (ASV) table was created using the DADA2 pipeline in R and taxonomy assignment was performed using the SILVA database. Average sequencing depth after filtering was 138,323 and low sequencing counts were discarded based on 20% prevalence. Data were analyzed using the lmer and adonis procedures in RStudio ver. 4.2.2. Analysis of 16S rRNA genes did not detect pre- or post-weaning GOS effects on Chao1, Simpson, or Shannon alpha diversity measures (P &amp;gt; 0.1). Similarly, no pre-weaning GOS effects were detected in Principal Coordinates Analysis (PCoA) beta diversity plots (P &amp;gt; 0.1); however, pigs fed GOS post-weaning segregated from controls (P &amp;lt; 0.01). The relative abundance of several genera including Fusicatenbacter and Collinsella increased by 2.0 and 1.3-fold, respectively, in the cecum of NG+ pigs, whereas, Bacteroides decreased by 3.8-fold (FDR, P &amp;lt; 0.05). In the rectum, relative abundance of Corprococcus decreased by 2.0-fold and Fusicatenbacter increased by 1.5-fold in NG+ pigs compared with NG- pigs (FDR, P &amp;lt; 0.05). Collectively, these genera in the cecum and rectum comprise of less than 3% of the total community. In contrast, lactobacillus in the cecum (33%) and rectum (21%) had a large relative abundance but was unaffected by treatment. We conclude that supplementation of GOS in the post-weaning phase 1 nursery diet alters the hindgut microbial community and may contribute to improvements in growth performance.

Single-Cell RNA Sequencing Technology Landscape in 2023.

Single-cell RNA sequencing (scRNA-seq) has revolutionized our understanding of cellular heterogeneity and the dynamics of gene expression, bearing profound significance in stem cell research. Depending on the starting materials used for analysis, scRNA-seq encompasses scRNA-seq and single-nucleus RNA sequencing (snRNA-seq). scRNA-seq excels in capturing cellular heterogeneity and characterizing rare cell populations within complex tissues, while snRNA-seq is advantageous in situations where intact cell dissociation is challenging or undesirable (eg, epigenomic studies). A number of scRNA-seq technologies have been developed as of late, including but not limited to droplet-based, plate-based, hydrogel-based, and spatial transcriptomics. The number of cells, sequencing depth, and sequencing length in scRNA-seq can vary across different studies. Addressing current technical challenges will drive the future of scRNA-seq, leading to more comprehensive and precise insights into cellular biology and disease mechanisms informing therapeutic interventions.

Time of Origin and Environmental Drivers of Leukemia Immune Escape after Transplantation

CONCLUSIONS One of the key questions regarding cancer therapies is whether treatment-resistant subclones pre-exist therapy or originate at a later time-point from residual cancer cells that were kept under control until that time. Here, we addressed this issue in HLA loss leukemia post-transplantation relapses, representing one of the best characterized mechanisms of selective immune escape from the adoptive immune therapy of allogeneic hematopoietic cell transplantation (allo-HCT) ( Vago, NEJM, 2009). Serial purified acute myeloid leukemia (AML) samples from 11 different patients, collected before allo-HCT and at the time of the post-transplant HLA loss relapse, were subjected to high-depth (average 100x) whole genome sequencing (WGS). Patient and donor germline controls were obtained through in vitro expansion of T cells collected from patients before and after allo-HCT, respectively, and sequenced at an average 30x depth. Starting from the notion that mutations can be employed as molecular clocks to date the age of cells ( Alexandrov, Nature Genetics, 2015) and that mutational signatures recapitulate their exposure to mutagens ( Alexandrov, Nature, 2020), we combined WGS with Poisson-based modelling to perform subclonal deconvolution ( Caravagna, Bioinformatics, 2020), and derive the timing of the HLA loss rearrangement, intersecting results with the patient clinical history. The AML cell line OCI-AML3 was employed for validation experiments, assessing in vitro the ability of selected compounds to induce DNA damage (quantifying ϒH2Ax through immunofluorescence and flow cytometry) and specific mutational signatures (by the WGS of single cell clones, performed at an average depth of sequencing of 30x). By WGS data of longitudinal AML samples we successfully reconstructed the clonal dynamics leading to 12 HLA loss relapses, and in each subclone deconvolved mutational signatures. By applying Poisson-based timing models to homozygous (i.e. occurred before the copy neutral loss of hetherozygosity - cn-LOH - rearrangement) and heterozygous (i.e. occurred after the event) mutations in the rearranged region, in 10/12 HLA loss relapses we confidently dated the event to a post-transplant time-point (Fig. 1A). In all HLA loss subclones we could detect a C&amp;gt;A mutational signature that was absent in pre-transplant samples (Fig. 1B): combining literature data ( Kanter, Cell Stem Cell, 2021) and in in vitro experiments on the OCI-AML3 cell line, we reconducted this signature to the antiviral drug ganciclovir (GCV). By in vitro experiments, we showed that GCV induces DNA damage and homologous recombination repair in AML, an effect that could be possibly potentiated by the concomitant exposure to the immunosuppressant mycophenolate ( Fang, Genome Medicine, 2022). Concordantly, by retrospective analysis of 126 AML relapses after a family haploidentical transplantation occurred at our Institute between August 2002 and December 2020, we observed that the incidence of HLA loss was 41/95 (43.1%) in patients exposed to GCV and 5/31 (16.1%) in unexposed ones (OR: 3.94, 95%CI 1.39 to 11.16, p=0.009). Our data indicate that in the majority of patients HLA loss variants originate after transplant, and that the genotoxicity of ganciclovir is a potent facilitator of the phenomenon. Defining the time of origin of therapy resistance has fundamental clinical consequences, implying that immune escape clones can only be targeted by post-transplant interventions, and warranting the development of new tools to anticipate the rapid detection of relapse.

A Novel Comprehensive Tumor-Informed Plasma cfDNA Assay to Monitor Minimal Residual Disease for Hematological and Solid Malignancies

Background: Cell-free DNA (cfDNA) in peripheral blood plasma is a promising tool for monitoring minimal residual disease (MRD) in hematologic and solid malignancies. Unlike conventional biopsy methods, cfDNA extraction is non-invasive, and the short turnaround time enables the analysis of patient samples across multiple time points over an extended period to monitor disease progression. Currently available tests are difficult to scale because they detect only a limited number of mutations, require patient- and mutation-specific customization, and demand constant review by experts. Here, we describe a novel and comprehensive cfDNA testing platform encompassing all major clinically reliable genetic hotspots that we validated using plasma samples from a cancer patient cohort. We also analyzed plasma samples collected from a follicular lymphoma (FL) patient at multiple time points to illustrate disease relapse before clinical manifestation and to show the applicability of our assay in MRD testing. Methods: We constructed a panel of 216 clinically actionable genes, including exonic and intronic regions, capable of tracking point somatic mutations (SNVs/indels), structural variants, and microsatellite instability (MSI). To identify mutational events, we used deep sequencing (median exon coverage of at least 2000×) and a tumor-informed variant calling algorithm developed in-house. Next, we determined the limit of detection (LOD), sensitivity, and specificity of our assay using plasma samples from 74 patients with various solid cancer diagnoses. We tested our platform on standard reference dilutions (Seraseq® ctDNA Complete™ Mutation Mix 0.1%, 0.25%, 0.5%, 1%, 2.5%; Twist cfDNA Pan-cancer Reference Standard 0.1%, 0.25%, 0.5%, 1%, 2%) to validate our results. Our assay was validated by sequencing samples from the internal cohort (n = 74) in replicates and by repeat sequencing of the reference materials in various input amounts in triplicate. Finally, we analyzed plasma samples from an FL patient at five timepoints to showcase our ability to detect mutations in disease recurrence before clinical manifestation. Results: We established the capability of our assay using plasma samples from a cancer patient cohort (n = 74) and two sets of dilutions of reference materials. With an LOD &amp;gt; 0.05% variant allele frequency (VAF) and sequencing depth of more than 2000x our assay achieved a sensitivity of 93-97% and a specificity of 98-99%. It also detected a clonal hematopoiesis of indeterminate potential (CHIP) mutation in one plasma sample, implying that it was more sensitive than variant calling from solid tumors and may allow us to distinguish true somatic mutations from CHIP mutations. VAF comparison between two plasma replicates showed high correlation (Pearson's r &amp;gt; 0.99, p &amp;lt; 0.0001), indicating that our assay is robust and suitable for comparing samples from the same patient across multiple time points. Using different amounts of the reference cfDNA as the starting material for library preparation (15 ng, 25 ng, 50 ng), we assessed the variability of VAF to determine the error range associated with each amount used. By identifying the appropriate amount of starting material needed in order to maintain the LOD at 0.05% VAF across all samples analyzed, we were able to account for variations in cfDNA levels in plasma samples collected at different time points from the same patient. As such, it ensured the reliability of our test across all analyzed samples. We used our assay to analyze plasma samples from a patient diagnosed with stage III FL in 2014 who underwent six cycles of therapy with bendamustine and rituximab. The five plasma samples analyzed were collected in 2021, 2022, and 2023 ( Figure 1). Our assay detected clinically relevant mutations in samples from Mar 2021, which was 14 months earlier than mutation calling from blood cells. Conclusions: At a LOD of 0.05% VAF and with all performance metrics greater than 95%, our assay meets the requirements for MRD testing. Analysis of plasma samples from an FL patient also shows that our assay can detect clinically relevant mutations in plasma samples much earlier than mutation calling from blood cells. We will further validate our assay with more clinical samples and expand our scope to include structural variants and MSI, to further lower the LOD.

Analytical Validation of a Duplex Sequencing AML Assay with a Limit of Detection Below 0.01% VAF for SNVs and Indels

Sensitive, specific, precise, and accurate detection of low frequency mutations in acute myeloid leukemia (AML) is critical to assess measurable residual disease (MRD). Next generation sequencing (NGS) assays are powerful tools for this, permitting interrogation of thousands of base positions across genomic regions of interest. However, PCR and sequencing errors in these assays limit accurate mutation detection at variant allele frequencies (VAF) below approximately 1%. Duplex sequencing (DS), an error-corrected NGS (ecNGS) method, greatly reduces such errors by relating original top and bottom DNA strands to make double stranded consensus sequences, thus enabling sensitive, specific, accurate, and precise detection of variants below 0.01% VAF. Here, we report analytical validation of an updated DuplexSeq™ AML assay, a 36-gene panel informed by 2022 European LeukemiaNet (ELN) recommendations combined with kitted library preparation reagents and a cloud-based analysis pipeline. The assay uniformly targets regions that harbor AML-associated mutations: 82% of targeted regions exhibit sequencing depth &amp;gt;80% of the panel-wide mean and &amp;gt;99.5% of targeted regions exhibit depth &amp;gt;20% of the panel-wide mean. These analytical studies also validate an improved “v2 chemistry” version of a DS library preparation kit. This improved chemistry provides an enzymatic fragmentation step that replaces time-consuming mechanical fragmentation methods while increasing the yield of duplex data for a given input mass of DNA sample and eliminating damage that results from mechanical fragmentation. A contrived human genomic DNA sample carrying 26 variants targeted by the panel, DNA extracted from AML-positive peripheral blood (PB) and bone marrow aspirate (BMA) specimens, and DNA extracted from healthy normal PB specimens were used as test samples to execute limit of detection (LoD), limit of blank (LoB), accuracy, linearity, and repeatability &amp; reproducibility studies. In this assay a duplex depth threshold of ≥20,000x is applied; ≥91.8% of the 42,605 genomic base positions targeted by the panel are consistently (in 95% of analyzed libraries) above this threshold. The LoD study assayed samples carrying variants from 0.37% to 0.00055% VAF. The assay demonstrated excellent sensitivity, with variants down to 0.00055% VAF detected. A probit model was used to estimate the VAF at which 95% of variants would be detected, establishing an LoD of 0.008% VAF for single-nucleotide variants (SNVs) and insertions and deletions (indels) (see Figure 1). The LoB study assessed the background detection rate in DNA samples derived from healthy normal specimens. In &amp;gt;95% of sites surveyed, no variants were observed, establishing an assay LoB of 0 (0% VAF). Accuracy of the assay was determined by interrogating PB and BMA samples from AML-positive and -negative donors along with contrived test samples. Results were compared to data acquired via orthogonal methods. For variants above the LoD of 0.008% VAF, accuracy was high, with overall percent agreement (OPA) of 98.3%, positive percent agreement (PPA) of 100.0%, and negative percent agreement (NPA) of 98.3%. Data were linear from 0.0016% VAF to 96.1% VAF with R 2 = 0.95. All 6 expected FLT3-ITD variants, and all 5 expected NPM1 insertions were detected by the assay. Precision study data showed that the assay was highly repeatable, with OPA of 97.5%, PPA of 98.0%, and NPA of 97.3%. Reproducibility across 2 operators, 2 lots of reagents, and 3 independent runs of library preparation was also very high, with OPA of 97.5%, PPA of 98.8%, and NPA of 96.9%. Current ELN guidelines provisionally define NGS-MRD positivity as ≥0.1% VAF, but recent evidence suggests lower frequency variants may be informative. With an LoD of 0.008% VAF for SNVs and indels, and an LoB of 0, this novel AML assay comprising an updated gene panel coupled with improved duplex sequencing chemistry and software represents a highly sensitive, specific, accurate, and precise ecNGS assay for detecting variants associated with AML MRD.

Parallel Genomic Analysis from Paired Bone Marrow and Peripheral Blood Samples of 200 Cytopenic Patients

Background: Analysis of bone marrow is the gold standard in patients with unexplainable cytopenia for diagnosing (or excluding) myelodysplastic neoplasm (MDS). Besides morphology also molecular genetics and cytogenetics are currently assessed from BM as both are required for diagnosis and risk stratification. Substitution of BM analysis with peripheral blood (PB) analysis would be a substantial progress especially in the elderly to allow for a less invasive strategy. Aims: Validate, whether PB reliably reflects molecular and cytogenetic aberrations of BM by evaluating somatic mutations and copy number variations (CNVs) in concurrent PB and BM samples. Methods: We analyzed paired BM and PB samples (time difference of sampling: &amp;lt;15 days) of 200 cytopenic patients. All samples were subjected to targeted panel sequencing analyzing mutations (assessed as positive at variant allelic frequencies/VAFs &amp;gt;1%) in 40 genes associated with myeloid neoplasms (coverage: 1500x). If available, BM chromosome analysis (CA) and BM FISH analysis were compared to NGS CNV analysis from PB using a human CNV backbone hybridization panel (coverage: 50x). Results: In 69 patients (35%) no mutations (MUTs) were detected in either BM or PB (Figure 1A). For the remaining 131 patients at least one MUT was found in BM (median MUTs/patient: 2; range 1-10). In 130/131 patients MUTs were detected in both BM and PB (same number of MUTs: n=121; higher number of MUTs in BM than in PB: n=9) while in only one patient harboring a DNMT3A MUT as the sole MUT in BM (VAF: 9%) no MUT was detected in PB. This resulted in a negative predictive value of 98.6% (69/70). Overall, we detected 364 MUTs in 29 of the analyzed genes (BM MUTs: n=362; PB MUTs: n=349) and a concordance between BM and PB of 95% (347/364; Figure 1A). The median VAF was lower in PB compared to BM across all MUTs (median VAF: 20% vs. 12%; p&amp;lt;0.001). All MUTs detected in BM at VAF ≥13% (n=229) were also detected in corresponding PB. Next, we compared CNVs detected by PB NGS to BM karyotype (KT). For 119/200 patients PB NGS including a CNV-backbone panel was available. BM CA revealed a normal KT in 71 and an aberrant KT in 48 patients. In total, 126 aberrations were detected in BM CA including 7 balanced aberrations that can technically not be detected using the CNV backbone panel. CNV analysis from PB revealed 68 aberrations (of 119 detectable BM CNVs; 54%) with 64 being concordant with BM CA. In total, in 88/119 patients (74%) concordant results were found in BM and PB (70/71 with normal and 18/48 with aberrant BM KT). In 3 patients (2.5%) different CNVs were detected (1/3: no CNV in BM CA), while in 28 (23.5%) either fewer CNVs (n=8; all complex KTs) or no CNVs (n=20) were detected in PB compared to BM. The majority (76%) of BM CNVs not detected by PB NGS with available BM FISH data were small BM clones (clone size ≤30%) and therefore not reliably captured by NGS PB with a technical sensitivity of 30%. Interestingly, 4 of the discordant patients with aberrant BM KT did not harbor any somatic MUT in BM and PB. Regarding the IPSS-R cytogenetic risk group, 95 (80%) patients fell into the same category, while 22 (18%; all with aberrant BM KT) were assigned to a better risk category by NGS PB than based on BM CA data and 2 patients (2%) to a worse risk category by NGS PB (Figure 1B). Conclusion: In cytopenic patients mutations can be reliably detected from PB showing a concordance with BM of individual mutations of 95% and a consistency regarding overall mutation detection between PB and BM in 99.5% of patients. Cytogenetic risk group assignment based on NGS CNV analysis from PB compared to chromosome analysis from BM resulted in the same risk group in 80% of patients. However, to further decrease the number of discrepant cases, the technical limitations (sensitivity, inability to detect balanced CNVs) of the CNV backbone method might be overcome by increasing the sequencing depth and/or using CD34-enriched PB samples and adding probe sets. Alternatively, FISH on PB smears could be used for CNV detection. Overall, our data indicate a high degree of overlap between PB and BM regarding mutational and cytogenetic analysis in patients with unexplained cytopenia and suggest PB as reliable surrogate especially in patients without any genetic aberrations.

Clonal Hematopoiesis Is Associated with Cardiovascular Disease Risk and Worse Survival after Autologous Hematopoietic Cell Transplantation for Lymphoma

Introduction: Autologous hematopoietic cell transplantation (HCT) is an effective treatment for patients with relapsed or refractory lymphoma, and use of HCT has contributed to a growing number of survivors. Despite these improvements in outcomes, long-term survivors of lymphoma are at high risk of developing clinically significant cardiovascular disease (CVD) which, to date, has largely been attributed to pre-HCT cardiotoxic therapeutic exposures (e.g. anthracyclines, radiation) and modifiable risk factors (e.g. hypertension, diabetes, dyslipidemia) after HCT. In the general population, clonal hematopoiesis of indeterminate potential (CHIP), which refers to clonal expansion of hematopoietic stem cells with somatic mutations, has been associated with aging-related health conditions including CVD. Accumulating data suggests CHIP is more prevalent in cancer patients compared to age-matched controls, due in part to clonal selection after genotoxic stressors. There is a paucity of information about the association between CHIP and CVD after HCT, including its interaction with modifiable risk factors and subsequent impact on survival after HCT. Methods: This was a retrospective cohort study of 861 consecutive patients who underwent autologous HCT for lymphoma between 2010 and 2016, with available pre-HCT DNA obtained from peripheral blood stem cells. We performed targeted exome sequencing of DNA (92 pre-defined gene panel with 1000x sequencing depth) to determine the presence of CHIP (variant allele frequency ≥2%). CVD outcomes of interest included de novo heart failure, coronary artery disease and stroke after HCT. Five-year cumulative incidence of CVD was calculated taking into consideration competing risk of death. Fine-Gray multivariable regression was used to evaluate the association between CHIP and CVD, adjusting for demographic, clinical and therapeutic factors. The Kaplan-Meier method was used to examine the effect of CHIP on overall survival after HCT. Results: Median age at HCT was 55.7 years (range 18.4-78.1 years); 63.3% were male; race/ethnicity: non-Hispanic White (56.9%), Hispanic (26.1%), Asian (10.8%), Black (5.1%); 78.0% had non-Hodgkin lymphoma; comorbidities at HCT: hypertension (32.1%), dyslipidemia (16.4%), diabetes (10.1%), high (≥3) HCT-comorbidity index (CI) score (46.1%). Nearly all patients (94.3%) had received anthracycline-based chemotherapy prior to HCT, and 63.3% were treated with cumulative dose ≥300mg/m 2; 6% had chest radiotherapy. Overall, 186 patients (21.7% of the cohort) had at least one CHIP variant and 57 (6.6%) had ≥2 variants. DNMT3A, PPM1D, TET2 and TP53 were the most frequently mutated genes. In total, 74 patients developed CVD within five years from HCT. The 5-year cumulative incidence of CVD was significantly higher in patients with CHIP (20.9% vs 9.1%, p&amp;lt;0.001), compared to those without CHIP, and there was a graded relationship between number of CHIP variants and incidence of CVD. In the adjusted (age, cancer diagnosis, HCT-CI, modifiable cardiovascular risk factors) multivariable model, CHIP was significantly and independently associated with a nearly two-fold (HR: 1.8, 95% CI=1.1-3.1; P=0.020) risk of CVD, as well as the individual outcome of heart failure (HR: 2.3, 95%CI=1.3-4.3; P=0.008). Patients who had both CHIP and hypertension had a 3-fold (HR: 3.0, 95%CI=1.4-6.5, p=0.006) risk of CVD (reference: no CHIP, no hypertension). Overall 5-year survival after HCT was significantly worse in patients with CHIP, compared to those without CHIP (47.2% vs 68.2%, p&amp;lt;0.001; adjusted HR: 1.9, 95%CI=1.4-2.6), attributed to the higher burden of non-relapse mortality in patients with CHIP. Conclusions: In this well-characterized and demographically diverse cohort of lymphoma patients undergoing HCT, CHIP was highly prevalent and associated with a significantly higher risk of CVD after HCT, with the highest risk among those with CHIP and hypertension. We confirmed the poor overall survival rate in patients with CHIP, independent of patient and disease-related risk factors. These findings highlight the potential role of CHIP as a novel biomarker to better define CVD risk in patients with lymphoma prior to HCT, setting the stage for better screening and implementation of interventions (e.g., aggressive management of modifiable risk factors) to reduce the risk of CVD and to ultimately improve long-term health outcomes after HCT.