An aptamer-mediated base editing platform for simultaneous knock-in and multiple gene knockout for allogeneic CAR-T cells generation.

Gene editing technologies hold promise for enabling the next generation of adoptive cellular therapies. Conventional gene editing platforms that rely on nuclease activity, such as Clustered regularly interspaced short palindromic repeats-CRISPR associated protein 9 (CRISPR-Cas9), allow efficient introduction of genetic modifications; however, these modifications occur via the generation of DNA double-strand breaks (DSBs) and can lead to unwanted genomic alterations and genotoxicity. Here, we apply a novel modular RNA aptamer-mediated Pin-point™ base editing platform to simultaneously introduce multiple gene knockouts and site-specific integration of a transgene in human primary T cells. We demonstrate high editing efficiency and purity at all target sites and significantly reduced frequency of chromosomal translocations compared to the conventional CRISPR-Cas9 system. Site-specific knock-in of a chimeric antigen receptor (CAR) and multiplex gene knockout are achieved within a single intervention and without the requirement for additional sequence-targeting components. The ability to perform complex genome editing efficiently and precisely highlights the potential of the Pin-point platform for application in a range of advanced cell therapies.

Evaluating reproducibility of AI algorithms in digital pathology with DAPPER.

Artificial Intelligence is exponentially increasing its impact on healthcare. As deep learning is mastering computer vision tasks, its application to digital pathology is natural, with the promise of aiding in routine reporting and standardizing results across trials. Deep learning features inferred from digital pathology scans can improve validity and robustness of current clinico-pathological features, up to identifying novel histological patterns, e.g., from tumor infiltrating lymphocytes. In this study, we examine the issue of evaluating accuracy of predictive models from deep learning features in digital pathology, as an hallmark of reproducibility. We introduce the DAPPER framework for validation based on a rigorous Data Analysis Plan derived from the FDA's MAQC project, designed to analyze causes of variability in predictive biomarkers. We apply the framework on models that identify tissue of origin on 787 Whole Slide Images from the Genotype-Tissue Expression (GTEx) project. We test three different deep learning architectures (VGG, ResNet, Inception) as feature extractors and three classifiers (a fully connected multilayer, Support Vector Machine and Random Forests) and work with four datasets (5, 10, 20 or 30 classes), for a total of 53, 000 tiles at 512 × 512 resolution. We analyze accuracy and feature stability of the machine learning classifiers, also demonstrating the need for diagnostic tests (e.g., random labels) to identify selection bias and risks for reproducibility. Further, we use the deep features from the VGG model from GTEx on the KIMIA24 dataset for identification of slide of origin (24 classes) to train a classifier on 1, 060 annotated tiles and validated on 265 unseen ones. The DAPPER software, including its deep learning pipeline and the Histological Imaging-Newsy Tiles (HINT) benchmark dataset derived from GTEx, is released as a basis for standardization and validation initiatives in AI for digital pathology.

A promoter-level mammalian expression atlas.

Regulated transcription controls the diversity, developmental pathways and spatial organization of the hundreds of cell types that make up a mammal. Using single-molecule cDNA sequencing, we mapped transcription start sites (TSSs) and their usage in human and mouse primary cells, cell lines and tissues to produce a comprehensive overview of mammalian gene expression across the human body. We find that few genes are truly 'housekeeping', whereas many mammalian promoters are composite entities composed of several closely separated TSSs, with independent cell-type-specific expression profiles. TSSs specific to different cell types evolve at different rates, whereas promoters of broadly expressed genes are the most conserved. Promoter-based expression analysis reveals key transcription factors defining cell states and links them to binding-site motifs. The functions of identified novel transcripts can be predicted by coexpression and sample ontology enrichment analyses. The functional annotation of the mammalian genome 5 (FANTOM5) project provides comprehensive expression profiles and functional annotation of mammalian cell-type-specific transcriptomes with wide applications in biomedical research.

The ISCB Student Council Internship Program: Expanding computational biology capacity worldwide.

Education and training are two essential ingredients for a successful career. On one hand, universities provide students a curriculum for specializing in one's field of study, and on the other, internships complement coursework and provide invaluable training experience for a fruitful career. Consequently, undergraduates and graduates are encouraged to undertake an internship during the course of their degree. The opportunity to explore one's research interests in the early stages of their education is important for students because it improves their skill set and gives their career a boost. In the long term, this helps to close the gap between skills and employability among students across the globe and balance the research capacity in the field of computational biology. However, training opportunities are often scarce for computational biology students, particularly for those who reside in less-privileged regions. Aimed at helping students develop research and academic skills in computational biology and alleviating the divide across countries, the Student Council of the International Society for Computational Biology introduced its Internship Program in 2009. The Internship Program is committed to providing access to computational biology training, especially for students from developing regions, and improving competencies in the field. Here, we present how the Internship Program works and the impact of the internship opportunities so far, along with the challenges associated with this program.

Highlights from the Third International Society for Computational Biology (ISCB) European Student Council Symposium 2014.

In this meeting report, we give an overview of the talks, presentations and posters presented at the third European Symposium of the International Society for Computational Biology (ISCB) Student Council. The event was organized as a satellite meeting of the 13th European Conference for Computational Biology (ECCB) and took place in Strasbourg, France on September 6th, 2014.

A multi-omics dataset for the analysis of frontotemporal dementia genetic subtypes.

Understanding the molecular mechanisms underlying frontotemporal dementia (FTD) is essential for the development of successful therapies. Systematic studies on human post-mortem brain tissue of patients with genetic subtypes of FTD are currently lacking. The Risk and Modyfing Factors of Frontotemporal Dementia (RiMod-FTD) consortium therefore has generated a multi-omics dataset for genetic subtypes of FTD to identify common and distinct molecular mechanisms disturbed in disease. Here, we present multi-omics datasets generated from the frontal lobe of post-mortem human brain tissue from patients with mutations in MAPT, GRN and C9orf72 and healthy controls. This data resource consists of four datasets generated with different technologies to capture the transcriptome by RNA-seq, small RNA-seq, CAGE-seq, and methylation profiling. We show concrete examples on how to use the resulting data and confirm current knowledge about FTD and identify new processes for further investigation. This extensive multi-omics dataset holds great value to reveal new research avenues for this devastating disease.

A Type II-B Cas9 nuclease with minimized off-targets and reduced chromosomal translocations in vivo.

Streptococcus pyogenes Cas9 (SpCas9) and derived enzymes are widely used as genome editors, but their promiscuous nuclease activity often induces undesired mutations and chromosomal rearrangements. Several strategies for mapping off-target effects have emerged, but they suffer from limited sensitivity. To increase the detection sensitivity, we develop an off-target assessment workflow that uses Duplex Sequencing. The strategy increases sensitivity by one order of magnitude, identifying previously unknown SpCas9's off-target mutations in the humanized PCSK9 mouse model. To reduce off-target risks, we perform a bioinformatic search and identify a high-fidelity Cas9 variant of the II-B subfamily from Parasutterella secunda (PsCas9). PsCas9 shows improved specificity as compared to SpCas9 across multiple tested sites, both in vitro and in vivo, including the PCSK9 site. In the future, while PsCas9 will offer an alternative to SpCas9 for research and clinical use, the Duplex Sequencing workflow will enable a more sensitive assessment of Cas9 editing outcomes.

The Role of Knockout Olfactory Receptor Genes in Odor Discrimination.

To date, little is known about the role of olfactory receptor (OR) genes on smell performance. Thanks to the availability of whole-genome sequencing data of 802 samples, we identified 41 knockout (KO) OR genes (i.e., carriers of Loss of Function variants) and evaluated their effect on odor discrimination in 218 Italian individuals through recursive partitioning analysis. Furthermore, we checked the expression of these genes in human and mouse tissues using publicly available data and the presence of organ-related diseases in human KO (HKO) individuals for OR expressed in non-olfactory tissues (Fisher test). The recursive partitioning analysis showed that age and the high number (burden) of OR-KO genes impact the worsening of odor discrimination (p-value < 0.05). Human expression data showed that 33/41 OR genes are expressed in the olfactory system (OS) and 27 in other tissues. Sixty putative mouse homologs of the 41 humans ORs have been identified, 58 of which are expressed in the OS and 37 in other tissues. No association between OR-KO individuals and pathologies has been detected. In conclusion, our work highlights the role of the burden of OR-KO genes in worse odor discrimination.

Natural human knockouts and Mendelian disorders: deep phenotyping in Italian isolates.

Whole genome sequencing (WGS) allows the identification of human knockouts (HKOs), individuals in whom loss of function (LoF) variants disrupt both alleles of a given gene. HKOs are a valuable model for understanding the consequences of genes function loss. Naturally occurring biallelic LoF variants tend to be significantly enriched in "genetic isolates," making these populations specifically suited for HKO studies. In this work, a meticulous WGS data analysis combined with an in-depth phenotypic assessment of 947 individuals from three Italian genetic isolates led to the identification of ten biallelic LoF variants in ten OMIM genes associated with known autosomal recessive diseases. Notably, only a minority of the identified HKOs (C7, F12, and GPR68 genes) displayed the expected phenotype. For most of the genes, instead, (ACADSB, FANCL, GRK1, LGI4, MPO, PGAM2, and RP1L1), the carriers showed none or few of the signs and symptoms typically associated with the related diseases. Of particular interest is a case presenting with a FANCL biallelic LoF variant and a positive diepoxybutane test but lacking a full Fanconi anemia phenotypic spectrum. Identifying KO subjects displaying expected phenotypes suggests that the lack of correct genetic diagnoses may lead to inappropriate and delayed treatment. In contrast, the presence of HKOs with phenotypes deviating from the expected patterns underlines how LoF variants may be responsible for broader phenotypic spectra. Overall, these results highlight the importance of in-depth phenotypical characterization to understand the role of LoF variants and the advantage of studying these variants in genetic isolates.

Hearing Function: Identification of New Candidate Genes Further Explaining the Complexity of This Sensory Ability.

To date, the knowledge of the genetic determinants behind the modulation of hearing ability is relatively limited. To investigate this trait, we performed Genome-Wide Association Study (GWAS) meta-analysis using genotype and audiometric data (hearing thresholds at 0.25, 0.5, 1, 2, 4, and 8 kHz, and pure-tone averages of thresholds at low, medium, and high frequencies) collected in nine cohorts from Europe, South-Eastern USA, Caucasus, and Central Asia, for an overall number of ~9000 subjects. Three hundred seventy-five genes across all nine analyses were tagged by single nucleotide polymorphisms (SNPs) reaching a suggestive p-value (p < 10-5). Amongst these, 15 were successfully replicated using a gene-based approach in the independent Italian Salus in the Apulia cohort (n = 1774) at the nominal significance threshold (p < 0.05). In addition, the expression level of the replicated genes was assessed in published human and mouse inner ear datasets. Considering expression patterns in humans and mice, eleven genes were considered particularly promising candidates for the hearing function: BNIP3L, ELP5, MAP3K20, MATN2, MTMR7, MYO1E, PCNT, R3HDM1, SLC9A9, TGFB2, and YTHDC2. These findings represent a further contribution to our understanding of the genetic basis of hearing function and its related diseases.

A verified genomic reference sample for assessing performance of cancer panels detecting small variants of low allele frequency.

BACKGROUND: Oncopanel genomic testing, which identifies important somatic variants, is increasingly common in medical practice and especially in clinical trials. Currently, there is a paucity of reliable genomic reference samples having a suitably large number of pre-identified variants for properly assessing oncopanel assay analytical quality and performance. The FDA-led Sequencing and Quality Control Phase 2 (SEQC2) consortium analyze ten diverse cancer cell lines individually and their pool, termed Sample A, to develop a reference sample with suitably large numbers of coding positions with known (variant) positives and negatives for properly evaluating oncopanel analytical performance. RESULTS: In reference Sample A, we identify more than 40,000 variants down to 1% allele frequency with more than 25,000 variants having less than 20% allele frequency with 1653 variants in COSMIC-related genes. This is 5-100× more than existing commercially available samples. We also identify an unprecedented number of negative positions in coding regions, allowing statistical rigor in assessing limit-of-detection, sensitivity, and precision. Over 300 loci are randomly selected and independently verified via droplet digital PCR with 100% concordance. Agilent normal reference Sample B can be admixed with Sample A to create new samples with a similar number of known variants at much lower allele frequency than what exists in Sample A natively, including known variants having allele frequency of 0.02%, a range suitable for assessing liquid biopsy panels. CONCLUSION: These new reference samples and their admixtures provide superior capability for performing oncopanel quality control, analytical accuracy, and validation for small to large oncopanels and liquid biopsy assays.

Predictability of drug-induced liver injury by machine learning.

BACKGROUND: Drug-induced liver injury (DILI) is a major concern in drug development, as hepatotoxicity may not be apparent at early stages but can lead to life threatening consequences. The ability to predict DILI from in vitro data would be a crucial advantage. In 2018, the Critical Assessment Massive Data Analysis group proposed the CMap Drug Safety challenge focusing on DILI prediction. METHODS AND RESULTS: The challenge data included Affymetrix GeneChip expression profiles for the two cancer cell lines MCF7 and PC3 treated with 276 drug compounds and empty vehicles. Binary DILI labeling and a recommended train/test split for the development of predictive classification approaches were also provided. We devised three deep learning architectures for DILI prediction on the challenge data and compared them to random forest and multi-layer perceptron classifiers. On a subset of the data and for some of the models we additionally tested several strategies for balancing the two DILI classes and to identify alternative informative train/test splits. All the models were trained with the MAQC data analysis protocol (DAP), i.e., 10x5 cross-validation over the training set. In all the experiments, the classification performance in both cross-validation and external validation gave Matthews correlation coefficient (MCC) values below 0.2. We observed minimal differences between the two cell lines. Notably, deep learning approaches did not give an advantage on the classification performance. DISCUSSION: We extensively tested multiple machine learning approaches for the DILI classification task obtaining poor to mediocre performance. The results suggest that the CMap expression data on the two cell lines MCF7 and PC3 are not sufficient for accurate DILI label prediction. REVIEWERS: This article was reviewed by Maciej Kandula and Pawel P. Labaj.

Integrative Network Fusion: A Multi-Omics Approach in Molecular Profiling.

Recent technological advances and international efforts, such as The Cancer Genome Atlas (TCGA), have made available several pan-cancer datasets encompassing multiple omics layers with detailed clinical information in large collection of samples. The need has thus arisen for the development of computational methods aimed at improving cancer subtyping and biomarker identification from multi-modal data. Here we apply the Integrative Network Fusion (INF) pipeline, which combines multiple omics layers exploiting Similarity Network Fusion (SNF) within a machine learning predictive framework. INF includes a feature ranking scheme (rSNF) on SNF-integrated features, used by a classifier over juxtaposed multi-omics features (juXT). In particular, we show instances of INF implementing Random Forest (RF) and linear Support Vector Machine (LSVM) as the classifier, and two baseline RF and LSVM models are also trained on juXT. A compact RF model, called rSNFi, trained on the intersection of top-ranked biomarkers from the two approaches juXT and rSNF is finally derived. All the classifiers are run in a 10x5-fold cross-validation schema to warrant reproducibility, following the guidelines for an unbiased Data Analysis Plan by the US FDA-led initiatives MAQC/SEQC. INF is demonstrated on four classification tasks on three multi-modal TCGA oncogenomics datasets. Gene expression, protein expression and copy number variants are used to predict estrogen receptor status (BRCA-ER, N = 381) and breast invasive carcinoma subtypes (BRCA-subtypes, N = 305), while gene expression, miRNA expression and methylation data is used as predictor layers for acute myeloid leukemia and renal clear cell carcinoma survival (AML-OS, N = 157; KIRC-OS, N = 181). In test, INF achieved similar Matthews Correlation Coefficient (MCC) values and 97% to 83% smaller feature sizes (FS), compared with juXT for BRCA-ER (MCC: 0.83 vs. 0.80; FS: 56 vs. 1801) and BRCA-subtypes (0.84 vs. 0.80; 302 vs. 1801), improving KIRC-OS performance (0.38 vs. 0.31; 111 vs. 2319). INF predictions are generally more accurate in test than one-dimensional omics models, with smaller signatures too, where transcriptomics consistently play the leading role. Overall, the INF framework effectively integrates multiple data levels in oncogenomics classification tasks, improving over the performance of single layers alone and naive juxtaposition, and provides compact signature sizes.

Multi-omics integration for neuroblastoma clinical endpoint prediction.

BACKGROUND: High-throughput methodologies such as microarrays and next-generation sequencing are routinely used in cancer research, generating complex data at different omics layers. The effective integration of omics data could provide a broader insight into the mechanisms of cancer biology, helping researchers and clinicians to develop personalized therapies. RESULTS: In the context of CAMDA 2017 Neuroblastoma Data Integration challenge, we explore the use of Integrative Network Fusion (INF), a bioinformatics framework combining a similarity network fusion with machine learning for the integration of multiple omics data. We apply the INF framework for the prediction of neuroblastoma patient outcome, integrating RNA-Seq, microarray and array comparative genomic hybridization data. We additionally explore the use of autoencoders as a method to integrate microarray expression and copy number data. CONCLUSIONS: The INF method is effective for the integration of multiple data sources providing compact feature signatures for patient classification with performances comparable to other methods. Latent space representation of the integrated data provided by the autoencoder approach gives promising results, both by improving classification on survival endpoints and by providing means to discover two groups of patients characterized by distinct overall survival (OS) curves. REVIEWERS: This article was reviewed by Djork-Arné Clevert and Tieliu Shi.

Author Correction: Transcription start site profiling of 15 anatomical regions of the Macaca mulatta central nervous system.

The authors regret that Luba M. Pardo was omitted in error from the author list of the original version of this Data Descriptor. This omission has now been corrected in the HTML and PDF versions. The authors also regret that Anemieke Rozemuller was omitted in error from the Acknowledgements of the original version of this Data Descriptor. This omission has now been corrected in the HTML and PDF versions.

Neuronal hemoglobin affects dopaminergic cells' response to stress.

Hemoglobin (Hb) is the major protein in erythrocytes and carries oxygen (O2) throughout the body. Recently, Hb has been found synthesized in atypical sites, including the brain. Hb is highly expressed in A9 dopaminergic (DA) neurons of the substantia nigra (SN), whose selective degeneration leads to Parkinson's disease (PD). Here we show that Hb confers DA cells' susceptibility to 1-methyl-4-phenylpyridinium (MPP+) and rotenone, neurochemical cellular models of PD. The toxic property of Hb does not depend on O2 binding and is associated with insoluble aggregate formation in the nucleolus. Neurochemical stress induces epigenetic modifications, nucleolar alterations and autophagy inhibition that depend on Hb expression. When adeno-associated viruses carrying α- and ß-chains of Hb are stereotaxically injected into mouse SN, Hb forms aggregates and causes motor learning impairment. These results position Hb as a potential player in DA cells' homeostasis and dysfunction in PD.

Transcription start site profiling of 15 anatomical regions of the Macaca mulatta central nervous system.

Rhesus macaque was the second non-human primate whose genome has been fully sequenced and is one of the most used model organisms to study human biology and disease, thanks to the close evolutionary relationship between the two species. But compared to human, where several previously unknown RNAs have been uncovered, the macaque transcriptome is less studied. Publicly available RNA expression resources for macaque are limited, even for brain, which is highly relevant to study human cognitive abilities. In an effort to complement those resources, FANTOM5 profiled 15 distinct anatomical regions of the aged macaque central nervous system using Cap Analysis of Gene Expression, a high-resolution, annotation-independent technology that allows monitoring of transcription initiation events with high accuracy. We identified 25,869 CAGE peaks, representing bona fide promoters. For each peak we provide detailed annotation, expanding the landscape of 'known' macaque genes, and we show concrete examples on how to use the resulting data. We believe this data represents a useful resource to understand the central nervous system in macaque.

C9orf72 is differentially expressed in the central nervous system and myeloid cells and consistently reduced in C9orf72, MAPT and GRN mutation carriers.

A non-coding hexanucleotide repeat expansion (HRE) in C9orf72 is a common cause of amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD) acting through a loss of function mechanism due to haploinsufficiency of C9orf72 or a gain of function mediated by aggregates of bidirectionally transcribed HRE-RNAs translated into di-peptide repeat (DPR) proteins. To fully understand regulation of C9orf72 expression we surveyed the C9orf72 locus using Cap Analysis of Gene Expression sequence data (CAGEseq). We observed C9orf72 was generally lowly expressed with the exception of a subset of myeloid cells, particularly CD14+ monocytes that showed up to seven fold higher expression as compared to central nervous system (CNS) and other tissues. The expression profile at the C9orf72 locus showed a complex architecture with differential expression of the transcription start sites (TSSs) for the annotated C9orf72 transcripts between myeloid and CNS tissues suggesting cell and/or tissue specific functions. We further detected novel TSSs in both the sense and antisense strand at the C9orf72 locus and confirmed their existence in brain tissues and CD14+ monocytes. Interestingly, our experiments showed a consistent decrease of C9orf72 coding transcripts not only in brain tissue and monocytes from C9orf72-HRE patients, but also in brains from MAPT and GRN mutation carriers together with an increase in antisense transcripts suggesting these could play a role in regulation of C9orf72. We found that the non-HRE related expression changes cannot be explained by promoter methylation but by the presence of the C9orf72-HRE risk haplotype and unknown functional interactions between C9orf72, MAPT and GRN.

Comprehensive promoter level expression quantitative trait loci analysis of the human frontal lobe.

BACKGROUND: Expression quantitative trait loci (eQTL) analysis is a powerful method to detect correlations between gene expression and genomic variants and is widely used to interpret the biological mechanism underlying identified genome wide association studies (GWAS) risk loci. Numerous eQTL studies have been performed on different cell types and tissues of which the majority has been based on microarray technology. METHODS: We present here an eQTL analysis based on cap analysis gene expression sequencing (CAGEseq) data created from human postmortem frontal lobe tissue combined with genotypes obtained through genotyping arrays, exome sequencing, and CAGEseq. Using CAGEseq as an expression profiling technique combined with these different genotyping techniques allows measurement of the molecular effect of variants on individual transcription start sites and increases the resolution of eQTL analysis by also including the non-annotated parts of the genome. RESULTS: We identified 2410 eQTLs and show that non-coding transcripts are more likely to contain an eQTL than coding transcripts, in particular antisense transcripts. We provide evidence for how previously identified GWAS loci for schizophrenia (NRGN), Parkinson's disease, and Alzheimer's disease (PARK16 and MAPT loci) could increase the risk for disease at a molecular level. Furthermore, we demonstrate that CAGEseq improves eQTL analysis because variants obtained from CAGEseq are highly enriched for having a functional effect and thus are an efficient method towards the identification of causal variants. CONCLUSION: Our data contain both coding and non-coding transcripts and has the added value that we have identified eQTLs for variants directly adjacent to TSS. Future eQTL studies would benefit from combining CAGEseq with RNA sequencing for a more complete interpretation of the transcriptome and increased understanding of eQTL signals.

DGEclust: differential expression analysis of clustered count data.

We present a statistical methodology, DGEclust, for differential expression analysis of digital expression data. Our method treats differential expression as a form of clustering, thus unifying these two concepts. Furthermore, it simultaneously addresses the problem of how many clusters are supported by the data and uncertainty in parameter estimation. DGEclust successfully identifies differentially expressed genes under a number of different scenarios, maintaining a low error rate and an excellent control of its false discovery rate with reasonable computational requirements. It is formulated to perform particularly well on low-replicated data and be applicable to multi-group data. DGEclust is available at http://dvav.github.io/dgeclust/.