The shape of gene expression distributions matter: how incorporating distribution shape improves the interpretation of cancer transcriptomic data

Control of goal-directed and inflexible actions by dorsal striatal melanocortin systems, in coordination with the central nucleus of the amygdala

The dorsomedial striatum (DMS) is associated with flexible goal seeking, as opposed to routinized habits. Whether local mechanisms brake this function, for instance when habits may be adaptive, is incompletely understood. We find that a sub-population of dopamine D1 receptor-containing striatal neurons express the melanocortin-4 receptor (MC4R) for α-melanocyte stimulating hormone. These neurons within the DMS are necessary and sufficient for controlling the capacity of mice to flexibly adjust actions based on the likelihood that they will be rewarded. In investigating MC4R function, we found that it suppresses immediate-early gene levels in the DMS and concurrently, flexible goal seeking. MC4R+ neurons receive input from the central nucleus of the amygdala, and behavioral experiments indicate that they are functionally integrated into an amygdalo-striatal circuit that suppresses action flexibility in favor of routine. Publicly available spatial transcriptomics datasets were analyzed for gene transcript correlates of Mc4r expression across the striatal subregions, revealing considerable co-variation in dorsal structures. This insight led to the discovery that the function of MC4R in the dorsolateral striatum complements that in the DMS, in this case suppressing habit-like behavior. Altogether, our findings suggest that striatal MC4R controls the capacity for goal-directed and inflexible actions alike.

Changes in expression of VGF, SPECC1L, HLA-DRA and RANBP3L act with APOE E4 to alter risk for late onset Alzheimer's disease

While there are currently over 40 replicated genes with mapped risk alleles for Late Onset Alzheimer's disease (LOAD), the Apolipoprotein E locus E4 haplotype is still the biggest driver of risk, with odds ratios for neuropathologically confirmed E44 carriers exceeding 30 (95% confidence interval 16.59-58.75). We sought to address whether the APOE E4 haplotype modifies expression globally through networks of expression to increase LOAD risk. We have used the Human Brainome data to build expression networks comparing APOE E4 carriers to non-carriers using scalable mixed-datatypes Bayesian network (BN) modeling. We have found that VGF had the greatest explanatory weight. High expression of VGF is a protective signal, even on the background of APOE E4 alleles. LOAD risk signals, considering an APOE background, include high levels of SPECC1L, HLA-DRA and RANBP3L. Our findings nominate several new transcripts, taking a combined approach to network building including known LOAD risk loci.

High HER2 Intratumoral Heterogeneity Is a Predictive Factor for Poor Prognosis in Early-Stage and Locally Advanced HER2-Positive Breast Cancer

PURPOSE

Breast cancer tumors frequently have intratumoral heterogeneity (ITH). Tumors with high ITH cause therapeutic resistance and have human epidermal growth factor receptor 2 (HER2) heterogeneity in response to HER2-targeted therapies. This study aimed to investigate whether high HER2 heterogeneity levels were clinically related to a poor prognosis for HER2-targeted adjuvant therapy resistance in primary breast cancers.

METHODS

This study included patients with primary breast cancer (n = 251) treated with adjuvant HER2-targeted therapies. HER2 heterogeneity was manifested by the shape of HER2 fluorescence in situ hybridization amplification (FISH) distributed histograms with the HER2 gene copy number within a tumor sample. Each tumor was classified into a biphasic grade graph (high heterogeneity [HH]) group or a monophasic grade graph (low heterogeneity [LH]) group based on heterogeneity. Both groups were evaluated for disease-free survival (DFS) and overall survival (OS) for a median of ten years of annual follow-up.

RESULTS

Of 251 patients with HER2-positive breast cancer, 46 (18.3%) and 205 (81.7%) were classified into the HH and LH groups, respectively. The HH group had more distant metastases and a poorer prognosis than the LH group (DFS: p < 0.001 (HH:63% vs. LH:91% at 10 years) and for the OS: p = 0.012 (HH:78% vs. LH:95% at 10 years).

CONCLUSIONS

High HER2 heterogeneity is a poor prognostic factor in patients with HER2-positive breast cancer. A novel approach to heterogeneity, which is manifested by the shape of HER2 FISH distributions, might be clinically useful in the prognosis prediction of patients after HER2 adjuvant therapy.

Development of an m6A-Related lncRNAs Signature Predicts Tumor Stemness and Prognosis for Low-Grade Glioma Patients

Background

Growing evidence has revealed that m6A modification of long noncoding RNAs (lncRNAs) dynamically controls tumor stemness and tumorigenesis-related processes. However, the prognostic significance of m6A-related lncRNAs and their associations with stemness in low-grade glioma (LGG) remain to be clarified.

Methods

A multicenter transcriptome analysis of lncRNA expression in 1,247 LGG samples was performed in this study. The stemness landscape of LGG tumors was presented and associations with clinical features were revealed. The m6A-related lncRNAs were identified between stemness groups and were further prioritized via least absolute shrinkage and selection operator Cox regression analysis. A risk score model based on m6A-related lncRNAs was constructed and validated in external LGG datasets.

Results

Based on the expression of LINC02984, PFKP-DT, and CRNDE, a risk model and nomogram were constructed; they successfully predicted the survival of patients and were extended to external datasets. Significant correlations were observed between the risk score and tumor stemness. Moreover, patients in different risk groups exhibited distinct tumor immune microenvironments and immune signatures. We finally provided several potential compounds suitable for specific risk groups, which may aid in LGG treatment.

Conclusions

This novel signature presents noteworthy value in the prediction of prognosis and stemness status for LGG patients and will foster future research on the development of clinical regimens.

Batch correction of single-cell sequencing data via an autoencoder architecture

Motivation

Technical differences between gene expression sequencing experiments can cause variations in the data in the form of batch effect biases. These do not represent true biological variations between samples and can lead to false conclusions or hinder the ability to integrate multiple datasets. Since there is a growing need for the joint analysis of single-cell sequencing datasets from different sources, there is also a need to correct the resulting batch effects while maintaining the true biological variations in the data.

Results

We developed a semi-supervised deep learning architecture called Autoencoder-based Batch Correction (ABC) for integrating single-cell sequencing datasets. Our method removes batch effects through a guided process of data compression using supervised cell type classifier branches for biological signal retention. It aligns the different batches using an adversarial training approach. We comprehensively evaluate the performance of our method using four single-cell sequencing datasets and multiple measures for batch effect removal and biological variation conservation. ABC outperforms 10 state-of-the-art methods for this task including Seurat, scGen, ComBat, scanorama, scVI, scANVI, AutoClass, Harmony, scDREAMER, and CLEAR, correcting various types of batch effects while preserving intricate biological variations.

The Human Brainome: changes in expression of VGF, SPECC1L, HLA-DRA and RANBP3L act with APOE E4 to alter risk for late onset Alzheimer's disease

While there are currently over 40 replicated genes with mapped risk alleles for Late Onset Alzheimer's disease (LOAD), the Apolipoprotein E locus E4 haplotype is still the biggest driver of risk, with odds ratios for neuropathologically confirmed E44 carriers exceeding 30 (95% confidence interval 16.59-58.75). We sought to address whether the APOE E4 haplotype modifies expression globally through networks of expression to increase LOAD risk. We have used the Human Brainome data to build expression networks comparing APOE E4 carriers to non-carriers using scalable mixed-datatypes Bayesian network (BN) modeling. We have found that VGF had the greatest explanatory weight. High expression of VGF is a protective signal, even on the background of APOE E4 alleles. LOAD risk signals, considering an APOE background, include high levels of SPECC1L, HLA-DRA and RANBP3L. Our findings nominate several new transcripts, taking a combined approach to network building including known LOAD risk loci.

Gene regulatory network reconstruction: harnessing the power of single-cell multi-omic data

Inferring gene regulatory networks (GRNs) is a fundamental challenge in biology that aims to unravel the complex relationships between genes and their regulators. Deciphering these networks plays a critical role in understanding the underlying regulatory crosstalk that drives many cellular processes and diseases. Recent advances in sequencing technology have led to the development of state-of-the-art GRN inference methods that exploit matched single-cell multi-omic data. By employing diverse mathematical and statistical methodologies, these methods aim to reconstruct more comprehensive and precise gene regulatory networks. In this review, we give a brief overview on the statistical and methodological foundations commonly used in GRN inference methods. We then compare and contrast the latest state-of-the-art GRN inference methods for single-cell matched multi-omics data, and discuss their assumptions, limitations and opportunities. Finally, we discuss the challenges and future directions that hold promise for further advancements in this rapidly developing field.

Long-read sequencing identifies a common transposition haplotype predisposing for CLCNKB deletions

BACKGROUND

Long-read sequencing is increasingly used to uncover structural variants in the human genome, both functionally neutral and deleterious. Structural variants occur more frequently in regions with a high homology or repetitive segments, and one rearrangement may predispose to additional events. Bartter syndrome type 3 (BS 3) is a monogenic tubulopathy caused by deleterious variants in the chloride channel gene CLCNKB, a high proportion of these being large gene deletions. Multiplex ligation-dependent probe amplification, the current diagnostic gold standard for this type of mutation, will indicate a simple homozygous gene deletion in biallelic deletion carriers. However, since the phenotypic spectrum of BS 3 is broad even among biallelic deletion carriers, we undertook a more detailed analysis of precise breakpoint regions and genomic structure.

METHODS

Structural variants in 32 BS 3 patients from 29 families and one BS4b patient with CLCNKB deletions were investigated using long-read and synthetic long-read sequencing, as well as targeted long-read sequencing approaches.

RESULTS

We report a ~3 kb duplication of 3'-UTR CLCNKB material transposed to the corresponding locus of the neighbouring CLCNKA gene, also found on ~50 % of alleles in healthy control individuals. This previously unknown common haplotype is significantly enriched in our cohort of patients with CLCNKB deletions (45 of 51 alleles with haplotype information, 2.2 kb and 3.0 kb transposition taken together, p=9.16×10-9). Breakpoint coordinates for the CLCNKB deletion were identifiable in 28 patients, with three being compound heterozygous. In total, eight different alleles were found, one of them a complex rearrangement with three breakpoint regions. Two patients had different CLCNKA/CLCNKB hybrid genes encoding a predicted CLCNKA/CLCNKB hybrid protein with likely residual function.

CONCLUSIONS

The presence of multiple different deletion alleles in our cohort suggests that large CLCNKB gene deletions originated from many independently recurring genomic events clustered in a few hot spots. The uncovered associated sequence transposition haplotype apparently predisposes to these additional events. The spectrum of CLCNKB deletion alleles is broader than expected and likely still incomplete, but represents an obvious candidate for future genotype/phenotype association studies. We suggest a sensitive and cost-efficient approach, consisting of indirect sequence capture and long-read sequencing, to analyse disease-relevant structural variant hotspots in general.

Rapid and Comprehensive Identification of Nontuberculous Mycobacteria

Next-generation sequencing is a powerful tool to accurately identify pathogens. The MinION sequencer is best suited for the rapid identification of bacterial species due to its real-time sequence output. In this chapter, we introduce a method to identify nontuberculous mycobacteria (NTM) in one sequencing analysis from culture isolates using the MinION sequencer. NTM disease is now recognized as a growing global health concern due to its increasing incidence and prevalence. There are over 200 NTM species, of which the major pathogens are further classified into many subspecies showing different antibiotic susceptibilities. Therefore, identifying the pathogens at the subspecies level of NTM is necessary to select an appropriate treatment regimen. The protocol described here includes DNA extraction by lysis using silica beads, library preparation, sequencing by the MinION sequencer, and analysis of multilocus sequence typing using the software "mlstverse" and enables rapid and comprehensive identification of 175 species of NTM at the subspecies level with high sensitivity and accuracy.

Neuron-Glia-Ratio-Like Approach Evidenced for Limited Variability and In-Aggregate Circadian Shifts in Cortical Cell-Specific Transcriptomes

Regardless of shifts in levels of individual transcripts, it remains elusive whether natural variability in cell-specific transcriptomes within the cerebral cortex is limited in aggregate. It is also unclear whether cortical cell-specific transcriptomes might change dynamically in absence of cell number changes. Total variation in neuron- and glia-specific in-aggregate transcriptomes could be identified in a model-free way via glia-neuron ratio approach, by univariate median-to-median ratios comparing integral levels of cell-specific transcripts within a tissue sample. When deleterious, regenerative or developmental events affecting cortical cell numbers were subtle, median-to-median ratios demonstrated within-group variability not exceeding <20-25% in most cases. These levels of total variability might be explained in part by limited (~5-10%) circadian and stress-induced shifts in cell-specific cortical transcriptomes. Relevant in-aggregate transcriptomic alterations were identified after shifts in cell numbers induced by well-validated deleterious events including ischemia, traumatic injury, microglia's activation/depletion or specific mutations. Cortical median-to-median ratios also follow naturally occurring changes in the numbers of excitatory, inhibitory neurons and glial cells during perinatal brain development. These findings characterize cortical cell-specific transcriptomes as subjects to circadian shifts and lifetime events, urging the importance of reporting full details on an origin of any transcriptomic sample collected in vivo.

Expression of type 1 cannabinoid receptor gene in bipolar disorder

BACKGROUND

The Endocannabinoid System (ECBs) may have a crucial role in bipolar disorder (BD). Previous reports have not detected abnormalities in the expression of the cannabinoid receptor gene CNR1, encoding for CB₁. However, we hypothesized that differentiating between mania and depression may uncover differences in CNR1 expression levels.

METHODS

We recruited 44 subjects with BD type I (BD-I), in mania (n = 22) and depression (n = 22) and 25 Healthy Controls (HC). CNR1 gene expression was analyzed using a quantitative real-time polymerase chain reaction from peripheral blood mononuclear cells. Data were analyzed using frequentist non-parametric and Bayesian approaches (generalized location-scale model based on lognormal and gamma distributions).

RESULTS

Using the frequentist non-parametric approach, the depression group had lower CNR1 expression compared to the mania group (p = 0.004). In addition, there was a negative correlation between CNR1 expression and Hamilton Depression Scale score (rho = -0.37; p = 0.007). Bayesian analyses further revealed that CNR1 expression in the mania group was higher and less variable than among HC (>95% probability), while CNR1 expression in the depression group was lower and more variable than among HC (100% probability).

LIMITATIONS

Lack of participants with bipolar disorder in the euthymic phase, lack of toxicology screening and evaluation of CNR1 variants.

CONCLUSION

CNR1 expression is higher and less variable in mania than in depression. It is highly probable that these differences also distinguish individuals in different illness phases from healthy controls. Future studies are needed to clarify the role of the endocannabinoid system in bipolar disorder.

Graph-Based Integration of Histone Modification Profiles

In this work, we introduce a similarity-network-based approach to explore the role of interacting single-cell histone modification signals in haematopoiesis—the process of differentiation of blood cells. Histones are proteins that provide structural support to chromosomes. They are subject to chemical modifications—acetylation or methylation—that affect the degree of accessibility of genes and, in turn, the formation of different phenotypes. The concentration of histone modifications can be modelled as a continuous signal, which can be used to build single-cell profiles. In the present work, the profiles of cell types involved in haematopoiesis are built based on all the major histone modifications (i.e., H3K27ac, H3K27me3, H3K36me3, H3K4me1, H3K4me3, H3K9me3) by counting the number of peaks in the modification signals; then, the profiles are used to compute modification-specific similarity networks among the considered phenotypes. As histone modifications come as interacting signals, we applied a similarity network fusion technique to integrate these networks in a unique graph, with the aim of studying the simultaneous effect of all the modifications for the determination of different phenotypes. The networks permit defining of a graph-cut-based separation score for evaluating the homogeneity of subgroups of cell types corresponding to the myeloid and lymphoid phenotypes in the classical representation of the haematopoietic tree. Resulting scores show that separation into myeloid and lymphoid phenotypes reflects the actual process of haematopoiesis.

Altered gene expression and PTSD symptom dimensions in World Trade Center responders

Despite experiencing a significant trauma, only a subset of World Trade Center (WTC) rescue and recovery workers developed posttraumatic stress disorder (PTSD). Identification of biomarkers is critical to the development of targeted interventions for treating disaster responders and potentially preventing the development of PTSD in this population. Analysis of gene expression from these individuals can help in identifying biomarkers of PTSD. We established a well-phenotyped sample of 371 WTC responders, recruited from a longitudinal WTC responder cohort using stratified random sampling, by obtaining blood, self-reported and clinical interview data. Using bulk RNA-sequencing from whole blood, we examined the association between gene expression and WTC-related PTSD symptom severity on (i) highest lifetime Clinician-Administered PTSD Scale (CAPS) score, (ii) past-month CAPS score, and (iii) PTSD symptom dimensions using a 5-factor model of re-experiencing, avoidance, emotional numbing, dysphoric arousal and anxious arousal symptoms. We corrected for sex, age, genotype-derived principal components and surrogate variables. Finally, we performed a meta-analysis with existing PTSD studies (total N = 1016), using case/control status as the predictor and correcting for these variables. We identified 66 genes significantly associated with total highest lifetime CAPS score (FDR-corrected p < 0.05), and 31 genes associated with total past-month CAPS score. Our more granular analyses of PTSD symptom dimensions identified additional genes that did not reach statistical significance in our analyses with total CAPS scores. In particular, we identified 82 genes significantly associated with lifetime anxious arousal symptoms. Several genes significantly associated with multiple PTSD symptom dimensions and total lifetime CAPS score (SERPINA1, RPS6KA1, and STAT3) have been previously associated with PTSD. Geneset enrichment of these findings has identified pathways significant in metabolism, immune signaling, other psychiatric disorders, neurological signaling, and cellular structure. Our meta-analysis revealed 10 genes that reached genome-wide significance, all of which were downregulated in cases compared to controls (CIRBP, TMSB10, FCGRT, CLIC1, RPS6KB2, HNRNPUL1, ALDOA, NACA, ZNF429 and COPE). Additionally, cellular deconvolution highlighted an enrichment in CD4 T cells and eosinophils in responders with PTSD compared to controls. The distinction in significant genes between total lifetime CAPS score and the anxious arousal symptom dimension of PTSD highlights a potential biological difference in the mechanism underlying the heterogeneity of the PTSD phenotype. Future studies should be clear about methods used to analyze PTSD status, as phenotypes based on PTSD symptom dimensions may yield different gene sets than combined CAPS score analysis. Potential biomarkers implicated from our meta-analysis may help improve therapeutic target development for PTSD.

An introduction to new robust linear and monotonic correlation coefficients

BACKGROUND

The most common measure of association between two continuous variables is the Pearson correlation (Maronna et al. in Safari an OMC. Robust statistics, 2019. https://login.proxy.bib.uottawa.ca/login?url=https://learning.oreilly.com/library/view/-/9781119214687/?ar&orpq&email=^u). When outliers are present, Pearson does not accurately measure association and robust measures are needed. This article introduces three new robust measures of correlation: Taba (T), TabWil (TW), and TabWil rank (TWR). The correlation estimators T and TW measure a linear association between two continuous or ordinal variables; whereas TWR measures a monotonic association. The robustness of these proposed measures in comparison with Pearson (P), Spearman (S), Quadrant (Q), Median (M), and Minimum Covariance Determinant (MCD) are examined through simulation. Taba distance is used to analyze genes, and statistical tests were used to identify those genes most significantly associated with Williams Syndrome (WS).

RESULTS

Based on the root mean square error (RMSE) and bias, the three proposed correlation measures are highly competitive when compared to classical measures such as P and S as well as robust measures such as Q, M, and MCD. Our findings indicate TBL2 was the most significant gene among patients diagnosed with WS and had the most significant reduction in gene expression level when compared with control (P value = 6.37E-05).

CONCLUSIONS

Overall, when the distribution is bivariate Log-Normal or bivariate Weibull, TWR performs best in terms of bias and T performs best with respect to RMSE. Under the Normal distribution, MCD performs well with respect to bias and RMSE; but TW, TWR, T, S, and P correlations were in close proximity. The identification of TBL2 may serve as a diagnostic tool for WS patients. A Taba R package has been developed and is available for use to perform all necessary computations for the proposed methods.

Mixed Distribution Models Based on Single-Cell RNA Sequencing Data

Progress in single-cell RNA sequencing (scRNA-seq) has yielded a lot of valuable data. Analysis of these data can provide a new perspective for studying the intratumoral heterogeneity and identifying gene markers. In this paper, the scRNA-seq data of colorectal cancer (CRC) are analyzed, and it is found that the shape of the gene expression difference (GED) data shows certain distribution regularity. To study the distribution regularity, mixed stable-normal distribution (MSND) model and mixed stable-exponential distribution (MSED) model are constructed to fit the GED data. And the estimated parameters of MSND and MSED are used to describe some characteristics of their distribution. Through the comparison of root mean square error and the chi-squared goodness of fit test, it is found that the fitting effect of MSED and MSND are both better than that of stable distribution and Cauchy distribution. Considering the given quantile thresholds, MSND and MSED can be used to identify tumor-related genes. The results of functional analysis indicate that the selected genes are highly correlated with CRC. In addition, the parameters of MSND and MSED exhibit a certain trend with the development of CRC. To explore the association, Gene-set enrichment analysis (GSEA) is performed. The results of GSEA reveal that the trend can well characterize the intratumoral heterogeneity of CRC. In addition, the application of MSED model on hepatocellular carcinoma shows that our model can analyze other cancers. Overall, MSND model and MSED model can well fit the GED data in different disease stages, the parameters of the two models can characterize the heterogeneity of CRC tumor cells, and the two models can be used to identify genes highly correlated with tumors.

Accelerating bioinformatics research with International Conference on Intelligent Biology and Medicine 2020

The International Association for Intelligent Biology and Medicine (IAIBM) is a nonprofit organization that promotes intelligent biology and medical science. It hosts an annual International Conference on Intelligent Biology and Medicine (ICIBM), which was initially established in 2012. Due to the coronavirus (COVID-19) pandemic, the ICIBM 2020 was held for the first time as a virtual online conference on August 9 to 10. The virtual conference had ~ 300 registered participants and featured 41 online real-time presentations. ICIBM 2020 received a total of 75 manuscript submissions, and 12 were selected to be published in this special issue of BMC Bioinformatics. These 12 manuscripts cover a wide range of bioinformatics topics including network analysis, imaging analysis, machine learning, gene expression analysis, and sequence analysis.