NetProphet 3: a machine learning framework for transcription factor network mapping and multi-omics integration.

MOTIVATION: Many methods have been proposed for mapping the targets of transcription factors (TFs) from gene expression data. It is known that combining outputs from multiple methods can improve performance. To date, outputs have been combined by using either simplistic formulae, such as geometric mean, or carefully hand-tuned formulae that may not generalize well to new inputs. Finally, the evaluation of accuracy has been challenging due to the lack of genome-scale, ground-truth networks. RESULTS: We developed NetProphet3, which combines scores from multiple analyses automatically, using a tree boosting algorithm trained on TF binding location data. We also developed three independent, genome-scale evaluation metrics. By these metrics, NetProphet3 is more accurate than other commonly used packages, including NetProphet 2.0, when gene expression data from direct TF perturbations are available. Furthermore, its integration mode can forge a consensus network from gene expression data and TF binding location data. AVAILABILITY AND IMPLEMENTATION: All data and code are available at https://zenodo.org/record/7504131#.Y7Wu3i-B2x8. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Dual threshold optimization and network inference reveal convergent evidence from TF binding locations and TF perturbation responses.

A high-confidence map of the direct, functional targets of each transcription factor (TF) requires convergent evidence from independent sources. Two significant sources of evidence are TF binding locations and the transcriptional responses to direct TF perturbations. Systematic data sets of both types exist for yeast and human, but they rarely converge on a common set of direct, functional targets for a TF. Even the few genes that are both bound and responsive may not be direct functional targets. Our analysis shows that when there are many nonfunctional binding sites and many indirect targets, nonfunctional sites are expected to occur in the cis-regulatory DNA of indirect targets by chance. To address this problem, we introduce dual threshold optimization (DTO), a new method for setting significance thresholds on binding and perturbation-response data, and show that it improves convergence. It also enables comparison of binding data to perturbation-response data that have been processed by network inference algorithms, which further improves convergence. The combination of dual threshold optimization and network inference greatly expands the high-confidence TF network map in both yeast and human. Next, we analyze a comprehensive new data set measuring the transcriptional response shortly after inducing overexpression of a yeast TF. We also present a new yeast binding location data set obtained by transposon calling cards and compare it to recent ChIP-exo data. These new data sets improve convergence and expand the high-confidence network synergistically.

SpDamID: Marking DNA Bound by Protein Complexes Identifies Notch-Dimer Responsive Enhancers.

We developed Split DamID (SpDamID), a protein complementation version of DamID, to mark genomic DNA bound in vivo by interacting or juxtapositioned transcription factors. Inactive halves of DAM (DNA adenine methyltransferase) were fused to protein pairs to be queried. Either direct interaction between proteins or proximity enabled DAM reconstitution and methylation of adenine in GATC. Inducible SpDamID was used to analyze Notch-mediated transcriptional activation. We demonstrate that Notch complexes label RBP sites broadly across the genome and show that a subset of these complexes that recruit MAML and p300 undergo changes in chromatin accessibility in response to Notch signaling. SpDamID differentiates between monomeric and dimeric binding, thereby allowing for identification of half-site motifs used by Notch dimers. Motif enrichment of Notch enhancers coupled with SpDamID reveals co-targeting of regulatory sequences by Notch and Runx1. SpDamID represents a sensitive and powerful tool that enables dynamic analysis of combinatorial protein-DNA transactions at a genome-wide level.

Inferring TF activities and activity regulators from gene expression data with constraints from TF perturbation data.

MOTIVATION: The activity of a transcription factor (TF) in a sample of cells is the extent to which it is exerting its regulatory potential. Many methods of inferring TF activity from gene expression data have been described, but due to the lack of appropriate large-scale datasets, systematic and objective validation has not been possible until now. RESULTS: We systematically evaluate and optimize the approach to TF activity inference in which a gene expression matrix is factored into a condition-independent matrix of control strengths and a condition-dependent matrix of TF activity levels. We find that expression data in which the activities of individual TFs have been perturbed are both necessary and sufficient for obtaining good performance. To a considerable extent, control strengths inferred using expression data from one growth condition carry over to other conditions, so the control strength matrices derived here can be used by others. Finally, we apply these methods to gain insight into the upstream factors that regulate the activities of yeast TFs Gcr2, Gln3, Gcn4 and Msn2. AVAILABILITY AND IMPLEMENTATION: Evaluation code and data are available at https://doi.org/10.5281/zenodo.4050573. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Past Roadblocks and New Opportunities in Transcription Factor Network Mapping.

One of the principal mechanisms by which cells differentiate and respond to changes in external signals or conditions is by changing the activity levels of transcription factors (TFs). This changes the transcription rates of target genes via the cell's TF network, which ultimately contributes to reconfiguring cellular state. Since microarrays provided our first window into global cellular state, computational biologists have eagerly attacked the problem of mapping TF networks, a key part of the cell's control circuitry. In retrospect, however, steady-state mRNA abundance levels were a poor substitute for TF activity levels and gene transcription rates. Likewise, mapping TF binding through chromatin immunoprecipitation proved less predictive of functional regulation and less amenable to systematic elucidation of complete networks than originally hoped. This review explains these roadblocks and the current, unprecedented blossoming of new experimental techniques built on second-generation sequencing, which hold out the promise of rapid progress in TF network mapping.

NetProphet 2.0: mapping transcription factor networks by exploiting scalable data resources.

MOTIVATION: Cells process information, in part, through transcription factor (TF) networks, which control the rates at which individual genes produce their products. A TF network map is a graph that indicates which TFs bind and directly regulate each gene. Previous work has described network mapping algorithms that rely exclusively on gene expression data and 'integrative' algorithms that exploit a wide range of data sources including chromatin immunoprecipitation sequencing (ChIP-seq) of many TFs, genome-wide chromatin marks, and binding specificities for many TFs determined in vitro. However, such resources are available only for a few major model systems and cannot be easily replicated for new organisms or cell types. RESULTS: We present NetProphet 2.0, a 'data light' algorithm for TF network mapping, and show that it is more accurate at identifying direct targets of TFs than other, similarly data light algorithms. In particular, it improves on the accuracy of NetProphet 1.0, which used only gene expression data, by exploiting three principles. First, combining multiple approaches to network mapping from expression data can improve accuracy relative to the constituent approaches. Second, TFs with similar DNA binding domains bind similar sets of target genes. Third, even a noisy, preliminary network map can be used to infer DNA binding specificities from promoter sequences and these inferred specificities can be used to further improve the accuracy of the network map. AVAILABILITY AND IMPLEMENTATION: Source code and comprehensive documentation are freely available at https://github.com/yiming-kang/NetProphet_2.0. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Model-based transcriptome engineering promotes a fermentative transcriptional state in yeast.

The ability to rationally manipulate the transcriptional states of cells would be of great use in medicine and bioengineering. We have developed an algorithm, NetSurgeon, which uses genome-wide gene-regulatory networks to identify interventions that force a cell toward a desired expression state. We first validated NetSurgeon extensively on existing datasets. Next, we used NetSurgeon to select transcription factor deletions aimed at improving ethanol production in Saccharomyces cerevisiae cultures that are catabolizing xylose. We reasoned that interventions that move the transcriptional state of cells using xylose toward that of cells producing large amounts of ethanol from glucose might improve xylose fermentation. Some of the interventions selected by NetSurgeon successfully promoted a fermentative transcriptional state in the absence of glucose, resulting in strains with a 2.7-fold increase in xylose import rates, a 4-fold improvement in xylose integration into central carbon metabolism, or a 1.3-fold increase in ethanol production rate. We conclude by presenting an integrated model of transcriptional regulation and metabolic flux that will enable future efforts aimed at improving xylose fermentation to prioritize functional regulators of central carbon metabolism.

Model-driven mapping of transcriptional networks reveals the circuitry and dynamics of virulence regulation.

Key steps in understanding a biological process include identifying genes that are involved and determining how they are regulated. We developed a novel method for identifying transcription factors (TFs) involved in a specific process and used it to map regulation of the key virulence factor of a deadly fungus-its capsule. The map, built from expression profiles of 41 TF mutants, includes 20 TFs not previously known to regulate virulence attributes. It also reveals a hierarchy comprising executive, midlevel, and "foreman" TFs. When grouped by temporal expression pattern, these TFs explain much of the transcriptional dynamics of capsule induction. Phenotypic analysis of TF deletion mutants revealed complex relationships among virulence factors and virulence in mice. These resources and analyses provide the first integrated, systems-level view of capsule regulation and biosynthesis. Our methods dramatically improve the efficiency with which transcriptional networks can be analyzed, making genomic approaches accessible to laboratories focused on specific physiological processes.

The developmental transcriptome of Drosophila melanogaster.

Drosophila melanogaster is one of the most well studied genetic model organisms; nonetheless, its genome still contains unannotated coding and non-coding genes, transcripts, exons and RNA editing sites. Full discovery and annotation are pre-requisites for understanding how the regulation of transcription, splicing and RNA editing directs the development of this complex organism. Here we used RNA-Seq, tiling microarrays and cDNA sequencing to explore the transcriptome in 30 distinct developmental stages. We identified 111,195 new elements, including thousands of genes, coding and non-coding transcripts, exons, splicing and editing events, and inferred protein isoforms that previously eluded discovery using established experimental, prediction and conservation-based approaches. These data substantially expand the number of known transcribed elements in the Drosophila genome and provide a high-resolution view of transcriptome dynamics throughout development.

Mapping functional transcription factor networks from gene expression data.

A critical step in understanding how a genome functions is determining which transcription factors (TFs) regulate each gene. Accordingly, extensive effort has been devoted to mapping TF networks. In Saccharomyces cerevisiae, protein-DNA interactions have been identified for most TFs by ChIP-chip, and expression profiling has been done on strains deleted for most TFs. These studies revealed that there is little overlap between the genes whose promoters are bound by a TF and those whose expression changes when the TF is deleted, leaving us without a definitive TF network for any eukaryote and without an efficient method for mapping functional TF networks. This paper describes NetProphet, a novel algorithm that improves the efficiency of network mapping from gene expression data. NetProphet exploits a fundamental observation about the nature of TF networks: The response to disrupting or overexpressing a TF is strongest on its direct targets and dissipates rapidly as it propagates through the network. Using S. cerevisiae data, we show that NetProphet can predict thousands of direct, functional regulatory interactions, using only gene expression data. The targets that NetProphet predicts for a TF are at least as likely to have sites matching the TF's binding specificity as the targets implicated by ChIP. Unlike most ChIP targets, the NetProphet targets also show evidence of functional regulation. This suggests a surprising conclusion: The best way to begin mapping direct, functional TF-promoter interactions may not be by measuring binding. We also show that NetProphet yields new insights into the functions of several yeast TFs, including a well-studied TF, Cbf1, and a completely unstudied TF, Eds1.

Cryptococcus neoformans dual GDP-mannose transporters and their role in biology and virulence.

Cryptococcus neoformans is an opportunistic yeast responsible for lethal meningoencephalitis in humans. This pathogen elaborates a polysaccharide capsule, which is its major virulence factor. Mannose constitutes over one-half of the capsule mass and is also extensively utilized in cell wall synthesis and in glycosylation of proteins and lipids. The activated mannose donor for most biosynthetic reactions, GDP-mannose, is made in the cytosol, although it is primarily consumed in secretory organelles. This compartmentalization necessitates specific transmembrane transporters to make the donor available for glycan synthesis. We previously identified two cryptococcal GDP-mannose transporters, Gmt1 and Gmt2. Biochemical studies of each protein expressed in Saccharomyces cerevisiae showed that both are functional, with similar kinetics and substrate specificities in vitro. We have now examined these proteins in vivo and demonstrate that cells lacking Gmt1 show significant phenotypic differences from those lacking Gmt2 in terms of growth, colony morphology, protein glycosylation, and capsule phenotypes. Some of these observations may be explained by differential expression of the two genes, but others suggest that the two proteins play overlapping but nonidentical roles in cryptococcal biology. Furthermore, gmt1 gmt2 double mutant cells, which are unexpectedly viable, exhibit severe defects in capsule synthesis and protein glycosylation and are avirulent in mouse models of cryptococcosis.

Role of fat body lipogenesis in protection against the effects of caloric overload in Drosophila.

The Drosophila fat body is a liver- and adipose-like tissue that stores fat and serves as a detoxifying and immune responsive organ. We have previously shown that a high sugar diet leads to elevated hemolymph glucose and systemic insulin resistance in developing larvae and adults. Here, we used stable isotope tracer feeding to demonstrate that rearing larvae on high sugar diets impaired the synthesis of esterified fatty acids from dietary glucose. Fat body lipid profiling revealed changes in both carbon chain length and degree of unsaturation of fatty acid substituents, particularly in stored triglycerides. We tested the role of the fat body in larval tolerance of caloric excess. Our experiments demonstrated that lipogenesis was necessary for animals to tolerate high sugar feeding as tissue-specific loss of orthologs of carbohydrate response element-binding protein or stearoyl-CoA desaturase 1 resulted in lethality on high sugar diets. By contrast, increasing the fat content of the fat body by knockdown of king-tubby was associated with reduced hyperglycemia and improved growth and tolerance of high sugar diets. Our work supports a critical role for the fat body and the Drosophila carbohydrate response element-binding protein ortholog in metabolic homeostasis in Drosophila.

The transcriptional diversity of 25 Drosophila cell lines.

Drosophila melanogaster cell lines are important resources for cell biologists. Here, we catalog the expression of exons, genes, and unannotated transcriptional signals for 25 lines. Unannotated transcription is substantial (typically 19% of euchromatic signal). Conservatively, we identify 1405 novel transcribed regions; 684 of these appear to be new exons of neighboring, often distant, genes. Sixty-four percent of genes are expressed detectably in at least one line, but only 21% are detected in all lines. Each cell line expresses, on average, 5885 genes, including a common set of 3109. Expression levels vary over several orders of magnitude. Major signaling pathways are well represented: most differentiation pathways are "off" and survival/growth pathways "on." Roughly 50% of the genes expressed by each line are not part of the common set, and these show considerable individuality. Thirty-one percent are expressed at a higher level in at least one cell line than in any single developmental stage, suggesting that each line is enriched for genes characteristic of small sets of cells. Most remarkable is that imaginal disc-derived lines can generally be assigned, on the basis of expression, to small territories within developing discs. These mappings reveal unexpected stability of even fine-grained spatial determination. No two cell lines show identical transcription factor expression. We conclude that each line has retained features of an individual founder cell superimposed on a common "cell line" gene expression pattern.

Steady progress and recent breakthroughs in the accuracy of automated genome annotation.

The sequencing of large, complex genomes has become routine, but understanding how sequences relate to biological function is less straightforward. Although much attention is focused on how to annotate genomic features such as developmental enhancers and non-coding RNAs, there is still no higher eukaryote for which we know the correct exon-intron structure of at least one ORF for each gene. Despite this uncomfortable truth, genome annotation has made remarkable progress since the first drafts of the human genome were analysed. By combining several computational and experimental methods, we are now closer to producing complete and accurate gene catalogues than ever before.

Multi-omics Integration Identifies Genes Influencing Traits Associated with Cardiovascular Risks: The Long Life Family Study.

The Long Life Family Study (LLFS) enrolled 4,953 participants in 539 pedigrees displaying exceptional longevity. To identify genetic mechanisms that affect cardiovascular risks in the LLFS population, we developed a multi-omics integration pipeline and applied it to 11 traits associated with cardiovascular risks. Using our pipeline, we aggregated gene-level statistics from rare-variant analysis, GWAS, and gene expression-trait association by Correlated Meta-Analysis (CMA). Across all traits, CMA identified 64 significant genes after Bonferroni correction (p ≤ 2.8×10-7), 29 of which replicated in the Framingham Heart Study (FHS) cohort. Notably, 20 of the 29 replicated genes do not have a previously known trait-associated variant in the GWAS Catalog within 50 kb. Thirteen modules in Protein-Protein Interaction (PPI) networks are significantly enriched in genes with low meta-analysis p-values for at least one trait, three of which are replicated in the FHS cohort. The functional annotation of genes in these modules showed a significant over-representation of trait-related biological processes including sterol transport, protein-lipid complex remodeling, and immune response regulation. Among major findings, our results suggest a role of triglyceride-associated and mast-cell functional genes FCER1A, MS4A2, GATA2, HDC, and HRH4 in atherosclerosis risks. Our findings also suggest that lower expression of ATG2A, a gene we found to be associated with BMI, may be both a cause and consequence of obesity. Finally, our results suggest that ENPP3 may play an intermediary role in triglyceride-induced inflammation. Our pipeline is freely available and implemented in the Nextflow workflow language, making it easily runnable on any compute platform (https://nf-co.re/omicsgenetraitassociation).

Whole Genome Linkage and Association Analyses Identify DLG Associated Protein-1 as a Novel Positional and Biological Candidate Gene for Muscle Strength: The Long Life Family Study.

BACKGROUND: Grip strength is a robust indicator of overall health, is moderately heritable, and predicts longevity in older adults. METHODS: Using genome-wide linkage analysis, we identified a novel locus on chromosome 18p (mega-basepair region: 3.4-4.0) linked to grip strength in 3 755 individuals from 582 families aged 64 ±â€…12 years (range 30-110 years; 55% women). There were 26 families that contributed to the linkage peak (cumulative logarithm of the odds [LOD] score = 10.94), with 6 families (119 individuals) accounting for most of the linkage signal (LOD = 6.4). In these 6 families, using whole genome sequencing data, we performed association analyses between the 7 312 single nucleotide (SNVs) and insertion deletion (INDELs) variants in the linkage region and grip strength. Models were adjusted for age, age2, sex, height, field center, and population substructure. RESULTS: We found significant associations between genetic variants (8 SNVs and 4 INDELs, p < 5 × 10-5) in the Disks Large-associated Protein 1 (DLGAP1) gene and grip strength. Haplotypes constructed using these variants explained up to 98.1% of the LOD score. Finally, RNAseq data showed that these variants were significantly associated with the expression of nearby Myosin Light Chain 12A (MYL12A), Structural Maintenance of Chromosomes Flexible Hinge Domain Containing 1 (SMCHD1), Erythrocyte Membrane Protein Band 4.1 Like 3 (EPB41L3) genes (p < .0004). CONCLUSIONS: The DLGAP1 gene plays an important role in the postsynaptic density of neurons; thus, it is both a novel positional and biological candidate gene for follow-up studies aimed at uncovering genetic determinants of muscle strength.

Discovery of genomic and transcriptomic pleiotropy between kidney function and soluble receptor for advanced glycation end products using correlated meta-analyses: The Long Life Family Study.

Patients with chronic kidney disease (CKD) have increased oxidative stress and chronic inflammation, which may escalate the production of advanced glycation end-products (AGEs). High soluble receptor for AGE (sRAGE) and low estimated glomerular filtration rate (eGFR) levels are associated with CKD and aging. We evaluated whether eGFR calculated from creatinine and cystatin C share pleiotropic genetic factors with sRAGE. We employed whole-genome sequencing and correlated meta-analyses on combined genome-wide association study (GWAS) p-values in 4182 individuals (age range: 24-110) from the Long Life Family Study (LLFS). We also conducted transcriptome-wide association studies (TWAS) on whole blood in a subset of 1209 individuals. We identified 59 pleiotropic GWAS loci (p < 5 × 10-8) and 17 TWAS genes (Bonferroni-p < 2.73 × 10-6) for eGFR traits and sRAGE. TWAS genes, LSP1 and MIR23AHG, were associated with eGFR and sRAGE located within GWAS loci, lncRNA-KCNQ1OT1 and CACNA1A/CCDC130, respectively. GWAS variants were eQTLs in the kidney glomeruli and tubules, and GWAS genes predicted kidney carcinoma. TWAS genes harbored eQTLs in the kidney, predicted kidney carcinoma, and connected enhancer-promoter variants with kidney function-related phenotypes at p < 5 × 10-8. Additionally, higher allele frequencies of protective variants for eGFR traits were detected in LLFS than in ALFA-Europeans and TOPMed, suggesting better kidney function in healthy-aging LLFS than in general populations. Integrating genomic annotation and transcriptional gene activity revealed the enrichment of genetic elements in kidney function and aging-related processes. The identified pleiotropic loci and gene expressions for eGFR and sRAGE suggest their underlying shared genetic effects and highlight their roles in kidney- and aging-related signaling pathways.

Toward an integrated model of capsule regulation in Cryptococcus neoformans.

Cryptococcus neoformans is an opportunistic fungal pathogen that causes serious human disease in immunocompromised populations. Its polysaccharide capsule is a key virulence factor which is regulated in response to growth conditions, becoming enlarged in the context of infection. We used microarray analysis of cells stimulated to form capsule over a range of growth conditions to identify a transcriptional signature associated with capsule enlargement. The signature contains 880 genes, is enriched for genes encoding known capsule regulators, and includes many uncharacterized sequences. One uncharacterized sequence encodes a novel regulator of capsule and of fungal virulence. This factor is a homolog of the yeast protein Ada2, a member of the Spt-Ada-Gcn5 Acetyltransferase (SAGA) complex that regulates transcription of stress response genes via histone acetylation. Consistent with this homology, the C. neoformans null mutant exhibits reduced histone H3 lysine 9 acetylation. It is also defective in response to a variety of stress conditions, demonstrating phenotypes that overlap with, but are not identical to, those of other fungi with altered SAGA complexes. The mutant also exhibits significant defects in sexual development and virulence. To establish the role of Ada2 in the broader network of capsule regulation we performed RNA-Seq on strains lacking either Ada2 or one of two other capsule regulators: Cir1 and Nrg1. Analysis of the results suggested that Ada2 functions downstream of both Cir1 and Nrg1 via components of the high osmolarity glycerol (HOG) pathway. To identify direct targets of Ada2, we performed ChIP-Seq analysis of histone acetylation in the Ada2 null mutant. These studies supported the role of Ada2 in the direct regulation of capsule and mating responses and suggested that it may also play a direct role in regulating capsule-independent antiphagocytic virulence factors. These results validate our experimental approach to dissecting capsule regulation and provide multiple targets for future investigation.

Evolution of genes and genomes on the Drosophila phylogeny.

Comparative analysis of multiple genomes in a phylogenetic framework dramatically improves the precision and sensitivity of evolutionary inference, producing more robust results than single-genome analyses can provide. The genomes of 12 Drosophila species, ten of which are presented here for the first time (sechellia, simulans, yakuba, erecta, ananassae, persimilis, willistoni, mojavensis, virilis and grimshawi), illustrate how rates and patterns of sequence divergence across taxa can illuminate evolutionary processes on a genomic scale. These genome sequences augment the formidable genetic tools that have made Drosophila melanogaster a pre-eminent model for animal genetics, and will further catalyse fundamental research on mechanisms of development, cell biology, genetics, disease, neurobiology, behaviour, physiology and evolution. Despite remarkable similarities among these Drosophila species, we identified many putatively non-neutral changes in protein-coding genes, non-coding RNA genes, and cis-regulatory regions. These may prove to underlie differences in the ecology and behaviour of these diverse species.