Furthermore, we collected a set of known regulators and targets [38] to test whether PROB could correctly reveal the regulatory causality. Over-expression plasmids and siRNA transfection. Since ψ0 is associated with the steady state and contains no dynamic information [26], we subtract the stationary component ψ0ψ0T from P, resulting in . Wilcoxon rank sum test (one-tailed) p value was calculated to assess the statistical significance. Investigation, The band intensity of western blots was detected by BLT GelView 6000M. Our proposed stage-weighted Gaussian kernel allows construction of diffusion-like random walks to quantify the temporal progression distance (TPD) between two patients (Eq (8)). (1). Algorithm for LADA diagnostic pathway based on autoantibody screening and C-peptide levels at diagnosis (to be used when financial restriction does not apply). Despite all this, PCR is being increasingly used for the rapid diagnosis of rapid. More specifically, for two trajectories s and , if the difference between the two inferred regulatory coefficients [aij(s)]n×n and is significantly larger than 0, then the difference between s and should not be arbitrarily small. https://doi.org/10.1371/journal.pcbi.1008379.t001. (b) R2 for the TCGA COAD dataset. Larger randomized studies are warranted to prove that DPP-4i might preserve C-peptide secretion. To elucidate the dynamic change of the EMT regulatory network during the progression, here, we applied PROB to an expression dataset of bladder cancer containing 84 UC samples and 28 SARC samples (GSE128192). The approval was based on data from phase III DEPICT clinical program (70). An edge was determined by examining whether the 95% credible interval (CI) of the parameter estimates did not contain zero (S4 Text). The faculty outlined some key points for future action, including a) screening for LADA, b) personalized medicine, c) need for more randomized controlled comparative trials with hypoglycemic agents, d) further investigation of immune therapy, e) large-scale long-term studies in different patient populations, f) quality of life/lifestyle issues, g) studies including patients of different ethnic origin, h) nature/quality of autoantibody assays (GADA, IA-2, etc. (f) Accuracy of GRN inference evaluated using area under curve (AUC) of ROC for the inferred network compared to the ground-truth network (AUC = 0.8395). The studies involving human participants were reviewed and approved by the Ethics Committee of Sun Yat-sen University Cancer Center (approval no. In a four-arm, randomized trial performed in 54 Chinese subjects, LADA patients were assigned to either sulfonylurea (SU) (n = 14) or rosiglitazone (n = 15) therapy if GADA was <175 units/mL and fasting C-peptide was >0.3 nmol/L. with , the data likelihood, given by Each sample was labeled with staging information (e.g., S1, S2, S3, and S4). (a) Kendall tau for the TCGA COAD dataset. Performance evaluation on real data indicates that PROB outperforms other methods in both pseudotime inference and GRN inference. (a-c) The expression levels of the above three genes in BT-20 breast cancer cells under FOXM1 siRNA or control (mock transfection and GFP siRNA) conditions were analyzed using a set of microarray data (GSE2222) [55] (d-f) RNA-seq data (GSE58626) [56] of MCF-7 breast cancer cells was used to analyze the differential expressions of the above three genes after FOXM1 inhibition by using small molecule compound IB that specifically inhibits FOXM1 [56]. The analysis showed that FOXM1 binds ASPM, CDCA8 and KIF2C in both cell lines (Fig 7C–7H). The levels of measurement variabilities in the synthetic data were quantified using the coefficient of variations (CVs) (from 0% to 30%). RT-PCR was performed using PrimeScript RT Master Mix (DRR036A, TakaRa) and qPCR was performed by qPCR SYBR Green Master Mix (11198ES03, Yeasen) in Real-time quantitative PCR instrument (Q1000+, Long Gene). (d) Recovered gene expression dynamics along with temporal progression. The statistical significance was assessed using Wilcoxon rank-sum test (one-tailed) p values. In a multicenter, randomized, nonblinded clinical study, Japanese patients with LADA, randomized to insulin or glibenclamide (n = 30 in each group), were followed for up to 5 years. The tumor tissues in this study were received from the operative resection of bladder cancer patients. To eliminate the effect of sampling density, we subsequently remold the above rotation-invariant kernel S(x, y) into an anisotropic kernel H(x, y), If patients are GADA positive, they are managed according to Fig. The use of metformin and/or insulin elicited much discussion among the panel, but it was concluded that they both have a role. The algorithm to infer progression trajectory and GRN is presented below. We assume that the conditional prior distribution of is the Laplace (double exponential) distribution with a mean of 0 and scale , that is, Denote the mean of the posterior distributions of aij as mij, and M = (mij)n×n. 40 hr post-transfection, the cells were plated on cover glass for recording. Citation: Sun X, Zhang J, Nie Q (2021) Inferring latent temporal progression and regulatory networks from cross-sectional transcriptomic data of cancer samples. A Japanese randomized trial comparing insulin (n = 30) with an SU (n = 30) over a 5-year period showed significantly better integrated C‐peptide response with insulin. received unrestricted educational grants from Novo Nordisk, Sanofi, MSD, and AstraZeneca. Only wild type cells (n = 479) without Stat1 and Ifnar1 knockout were chosen for analysis. A set of scRNA-seq data of dendritic cells stimulated with LPS was used for benchmarking. (e) PROB correctly revealed the ordering of the outgoing causality scores (on a log10 scale) for the known regulators and targets [38] on the DC scRNA-seq dataset. Is the Subject Area "Cancers and neoplasms" applicable to this article? The gene expression RNAseq and phenotype information associated with the TCGA COAD dataset were downloaded from the UCSC Xena website (https://xena.ucsc.edu/) via the following links: https://tcga-xena-hub.s3.us-east-1.amazonaws.com/latest/TCGA.COAD.sampleMap/HiSeqV2.gz and https://tcga-xena-hub.s3.us-east-1.amazonaws.com/latest/TCGA.COAD.sampleMap/COAD_clinicalMatrix; The gene expression RNAseq and phenotype information associated with the TCGA SKCM dataset were also downloaded from the UCSC Xena website (https://xena.ucsc.edu/) via the following links: https://tcga-xena-hub.s3.us-east-1.amazonaws.com/latest/TCGA.SKCM.sampleMap/HiSeqV2.gz and https://tcga-xena-hub.s3.us-east-1.amazonaws.com/latest/TCGA.SKCM.sampleMap/SKCM_clinicalMatrix. This model provides an area under the receiver operating characteristic curve (ROC AUC) of 0.90 (clinical features only) to 0.97 (all five parameters) with low prediction error, but it only applies to European-origin patients aged between 18 and 50 years at diagnosis and, being cross-sectional, is not predictive. Kaplan-Meier Plotter (http://kmplot.com) [67] was employed to perform analysis. The cells were sequenced at 1, 2, 4 and 6h after stimulation of LPS. To this end, a logistic regression model was developed to predict EMT states or histological subtypes (UC vs. SARC) of bladder cancer based on the expression levels of ACSS1 and PTPN12, which showed good predictive accuracy (S11 Fig). Acknowledgments. The temporal progression inference was performed to quantitatively order samples based on the whole gene expression profile with UC samples and SARC samples labeled by 1 and 2 respectively. The red lines represent the parameter values of aij used for generating the ground-truth network as in Equation (S22). ChIP-seq data were downloaded from GEO database (GSE62425) [54]. “Lady-like”: is there a latent autoimmune diabetes in the young? The permutation test p values for all three datasets were less than 0.05 (S12D–S12F Fig), verifying the non-randomness of the predicted targeted genes of FOXM1. However, not all patients in UKPDS and in other studies required insulin, even after 10 years from diagnosis. We next compared the accuracy of PROB with other existing GRN inference methods (e.g., PCOR, ARACNe, CLR, MRNET, Lasso, GENIE3, SCODE and LEAP) for cross-sectional data. Writing – review & editing, * E-mail: sunxq6@mail.sysu.edu.cn, xiaoqiangsun88@gmail.com (XS); qnie@uci.edu (QN), Affiliations Andersen et al. The knockdown or silence of FOXM1 significantly reduced the expressions of the above three genes. To reconstruct epithelial-to-mesenchymal transition (EMT) regulatory networks, we collected 44 representative genes of TGFB1 pathway, RhoA pathway, p53 pathway, p63 pathway and EMT transcriptional regulators (S1 Table) [39]. The authors concluded that in this small study, the primary outcome of safety was achieved, with evidence of a beneficial effect on β-cell function. (a-b) Root of mean squared error (RMSE) and Spearman correlation used for evaluating the accuracy of the temporal progression inference. Performance evaluation using both simulated data and real data demonstrate that our method outperforms other existing methods in both pseudotime inference and GRN inference. conceived the study, organized the meeting of the consensus group, agreed on the design of the algorithm, contributed to the definition of the guidelines, and wrote and approved the final draft of the manuscript. The risk of progression to insulin deficiency is variable, depending on age at diagnosis, autoantibody level, and presence of multiple islet autoantibodies (4,5,9,10,34). The robustness of the estimates of regulatory coefficients is justified through mathematical analysis and simulations. (10) Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. The Markov chain Monte Carlo (MCMC) algorithm with Gibbs sampling updates is employed to estimate the marginal distribution of each parameter. GADA is considered the most sensitive marker of LADA as it is the predominant autoantibody, whether in Europe or China, and in primary or in secondary care; e.g., the Action LADA study showed that approximately 90% of LADA subjects with diabetes-associated autoantibodies are GADA positive (9,15). Within LADA, C-peptide values, proxy for β-cell function, drive therapeutic decisions. Moreover, we used ChIP-seq data (GSE40762 [42]) to analyze the binding of FOXM1 to the predicted targeted genes (S7 Text). (18) The diffusion map, as a manifold-based nonlinear dimension reduction method, has been recently applied to scRNA-seq data analysis [26,57–59]. Yes We first define a Gaussian similarity function for two patients, x and y, as ACSS1 and PTPN12 have largest out-degree values in the UC-specific network and SARC-specific network, respectively. The numerical data underlying graphs in the manuscript is available at S1_Data.xlsx in the Supporting Information. The log-rank test p values for all three datasets were less than 1e-4. We approximate Conversely, an increase in C-peptide level was observed in those subjects treated with linagliptin; the difference between groups was significant at 28 and 58 weeks (P < 0.01 for all comparisons) (64). As such, the early changes in expressions of ACSS1 and PTPN12 during the progression of UC to SARC may be relevant for the early detection of SARC. (2) We compared PROB with its several variants: ‘ωxy = 1’ represents setting the weight coefficient ωxy in Eq (2) to be 1; ‘xref = random’ represents randomly assigning the reference point (Eq (9)) to identify the rooting point as the previous pseudotime inference methods usually did; ‘No stage’ represents leaving out the stage information, i.e., both ‘ωxy = 1’ and ‘xref = random’. Patients are thus ordered according to the TPD with respect to the root identified with the aid of staging information. participated in the meeting of the consensus group, agreed in the design of the LADA algorithm, critically discussed the format of the guidelines, critically revised the final draft of the manuscript, and approved the final manuscript. The panel concluded by advising general screening for LADA in newly diagnosed non–insulin-requiring diabetes and, importantly, that large randomized clinical trials are warranted. Among 8 targeted genes of FOXM1 predicted from PROB, 7 genes (including KIF2C, SHCBP1, CDCA8, NCAPG, ASPM, MELK and MCM10) were supported by the database information. by normalizing S with a proxy for the sampling density of the data points, LADA is a well-recognized form of diabetes; however, there are no guidelines for its management. To estimate the temporal order of cancer progression from the cross-sectional transcriptomic data, we develop a staging information-guided random walk approach to efficiently measure manifold distance between patients in a large cohort. Fig 4B and Fig 4C show the UC-specific network and the SARC-specific network, respectively, suggesting rewiring of the EMT regulatory network during the progression of UC to SARC. Investigation, The existing methods for inferring gene regulatory networks (GRNs) rely mainly on time-course gene expression data. 1–3). To address the above challenge, we propose to decode the latent temporal information underlying cancer progression via ordering patient samples based on transcriptomic similarity, and design a latent-temporal progression-based Bayesian method to infer GRNs from sample-based transcriptomic data of cancer patients. (b) Validation of the expression changes of the predicted target genes of FOXM1 with perturbation experiments. However, as expected, there was a reduced glycemic response to GLP-1RA analogs (exenatide/liraglutide) in a small patient group (n = 19) with diabetes-associated autoantibodies and low fasting C-peptide levels (≤0.25 nmol/L) (80). In contrast, cross-sectional studies (i.e., a snapshot of a particular group of people at a given point in time) based on high-throughput molecular omics data are more prevalent due to their relative feasibility. The identified ACSS1 is experimentally validated to promote epithelial-to-mesenchymal transition of bladder cancer cells, and the predicted FOXM1-targets interactions are verified and are predictive of relapse in breast cancer. However, many clinical laboratories have applied the mixed meal tolerance test values (87) for C-peptide measured 2 h after a random meal, although this has not been yet standardized. Both estrogen-dependent ER (+) MCF-7 and estrogen-independent ER (-) MDA-MB-231 human breast cancer cell lines were used for analysis. Management of Latent Autoimmune Diabetes in Adults: A Consensus Statement From an International Expert Panel, Immune abnormalities in diabetic patients not requiring insulin at diagnosis, Islet cell antibodies identify latent type I diabetes in patients aged 35-75 years at diagnosis, Antibodies to glutamic acid decarboxylase reveal latent autoimmune diabetes mellitus in adults with a non-insulin-dependent onset of disease, UKPDS 25: autoantibodies to islet-cell cytoplasm and glutamic acid decarboxylase for prediction of insulin requirement in type 2 diabetes, Clinical and genetic characteristics of type 2 diabetes with and without GAD antibodies, Latent autoimmune diabetes in adults (LADA), Non Insulin Requiring Autoimmune Diabetes Study Group, High titer of autoantibodies to GAD identifies a specific phenotype of adult-onset autoimmune diabetes, Number of autoantibodies and HLA genotype, more than high titers of glutamic acid decarboxylase autoantibodies, predict insulin dependence in latent autoimmune diabetes of adults, Adult-onset autoimmune diabetes in Europe is prevalent with a broad clinical phenotype: Action LADA 7, Frequency, immunogenetics, and clinical characteristics of latent autoimmune diabetes in China (LADA China study): a nationwide, multicenter, clinic-based cross-sectional study, Latent autoimmune diabetes in adults in the United Arab Emirates: clinical features and factors related to insulin-requirement, Etiology and pathogenesis of latent autoimmune diabetes in adults (LADA) compared to type 2 diabetes, Introducing the endotype concept to address the challenge of disease heterogeneity in type 1 diabetes, Islet cell autoimmunity in a triethnic adult population of the Third National Health and Nutrition Examination Survey, China National Diabetes and Metabolic Disorders Study Group, Identification of autoimmune type 1 diabetes and multiple organ-specific autoantibodies in adult-onset non-insulin-requiring diabetes in China: a population-based multicentre nationwide survey, A global perspective of latent autoimmune diabetes in adults, Type I diabetes masquerading as type II diabetes: possible implications for prevention and treatment, The many faces of diabetes: a disease with increasing heterogeneity, Adult-onset autoimmune diabetes: current knowledge and implications for management, The influence of type 2 diabetes-associated factors on type 1 diabetes. The capture time in the data was treated as an analogy to ‘staging’ information when using PROB. The gene expression data of 196 patients with clinical information (e.g., grade) were extracted from the GEO database (GSE7390 [40]). A previous study measured binding region coverage scores for 23 TFs and thus quantified their regulatory potential in the DCs using a high-throughput Chromatin ImmunoPrecipitation (HT-ChIP) method [37]. where E(x) is the row normalization of H, that is, Alum-formulated recombinant GADA (GAD-alum) was used in a small phase 2 study that was placebo-controlled with dose escalation in GADA-positive non–insulin-requiring patients (n = 47), who received subcutaneous injections of GAD-alum in different doses (81). The addition of other hypoglycemic agents such as incretin-based therapy (GLP-1RA or DPP-4i), TZD, and SGLT2i may confer some additional advantages, e.g., weight loss, cardiovascular/renal protection (Figs. Meanwhile, a wealth of time-course transcriptomic data has been generated from the laboratory experiments. First, to infer the latent temporal information of cancer progression from the cross-sectional data, a graph-based random walk approach was developed to quantitatively order patient samples (Fig 1A and 1B). The gene expression datasets as well as clinical information of the breast cancer patients used for network prediction were downloaded from the NCBI GEO database (GSE7390). An international group of experts convened a meeting to address this issue. For tumor sample-based gene expression data, several methods have been developed to infer gene networks. In this way, the cross-sectional data could be reordered to be analogous to time-course data. Conceptualization, In addition, we summarized and compared the capabilities of the above methods in predicting gene regulatory links, directions, signs and expression dynamics (Fig 3F). Funding: XS was supported by grants from the National Natural Science Foundation of China (11871070, 11931019), the Guangdong Basic and Applied Basic Research Foundation (2020B151502120), the Fundamental Research Funds for the Central Universities (20ykzd20). Asexuality is the lack of sexual attraction to others, or low or absent interest in or desire for sexual activity. C-peptide levels >0.7 nmol/L: suggests using a slightly modified ADA/EASD algorithm for T2D, notably the only difference being that LADA patients should be followed with repeated C-peptide measurements if there is a deterioration of glucose control; some of these cases will have false positive autoantibodies and therefore be true T2D. The studies involving human participants were reviewed and approved by the Ethics Committee of Sun Yat-sen University Cancer Center (approval no. https://doi.org/10.1371/journal.pcbi.1008379.s011. PROB provides a dynamic and systems perspective for characterizing and understanding cancer progression based on patients’ data. However, most available omics data from cross-sectional studies of cancer patients often lack sufficient temporal information, leading to a key challenge for GRN inference. No, Is the Subject Area "Genetic networks" applicable to this article? The differential expression of the above 8 genes between control condition and FOXM1 inhibition condition was examined to test whether they were down-regulated after FOXM1 inhibition. In sampling for C-peptide evaluation, the concomitant measurement of blood glucose levels should be done to ensure that it is between 80 and 180 mg/dL to avoid the effect of abnormally low or high glucose values. Temporal dynamics of gene expression (Fig 4D) showed that ACSS1 and PTPN12 oscillated synchronously with CDH1 (coding gene of epithelial marker protein E-cadherin) at the early stage of UC development. Other autoantibodies target different IA-2 epitopes (IA-2A), insulin (IAA), the islet-specific zinc transporter isoform 8 (ZnT8A), and tetraspanin 7, while other GADA epitopes are less frequent in LADA (15,16,37–41). The patients/participants provided their written informed consent to participate in this study. The statistical significance was assessed using Wilcoxon rank sum test (one-tailed) p values. The posterior distributions of the regulatory parameters against their true values show that the estimation was rather reliable (S3 Fig). To make a proposal for treatment of LADA, the panel reviewed current clinical trial data and reiterated the conclusions of the Cochrane Review regarding lack of good-quality, large-scale, controlled trials with long-term follow-up (54). Notably, a similar type of slowly progressive form of autoimmune diabetes can also be found in young-onset cases called latent autoimmune diabetes in the young (LADY) (21), reflecting a wide latitude for the variable age at diagnosis. A recent re-quantification of the LPS scRNA-seq dataset (GSE48968) was downloaded from the conquer database (http://imlspenticton.uzh.ch:3838/conquer/). Data obtained from all major studies including the UK Prospective Diabetes Study (UKPDS) (4) and the Botnia study (5) show that the autoantibody frequency (GADA) in patients diagnosed with T2D is higher in younger patients compared with older patients (e.g., in UKPDS from 34% when aged 25–34 years to 7% in older patients aged 55–65 years). C-peptide levels decrease more slowly in LADA than in T1D, and this marker may be used to stage LADA patients according to their residual β-cell function and progression toward insulin requirement (27,31–33). A subnetwork was reconstructed for FOXM1, which predicted that FOXM1 could positively regulate ASPM, CDCA8, KIF2C, MCM10, MELK, NCAPG, SHCBP1 and STIL (Fig 7A). Unraveling molecular regulatory networks underlying disease progression is critically important for understanding disease mechanisms and identifying drug targets. GZR2018-131). where are the random effects with each , (k = 0,1,⋯,m). No severe study-related adverse events occurred during the 5-year follow-up, and active treatment was not associated with increased risk of starting insulin treatment compared with placebo. The prior distribution of , is usually assumed to be an inverse gamma, with the probability distribution function Metformin can increase insulin sensitivity in T1D (55) without evidence that it could improve long-term glycemic control; in addition, it might reduce weight, LDL cholesterol levels, and the risk of atherosclerosis progression (56). The panel felt it was important to provide recommendations for the management of these patients in clinical practice. The Wilcoxon rank-sum test (one-tailed) p value was calculated to assess statistical significance. The qRT-PCR primer sequence used in this study are listed in S3 Table. The overall objective of a personalized approach for the management of LADA is to achieve good metabolic control and preserve β-cell function. The clinical gene expression data used for survival analysis were downloaded from the NCBI GEO database (GSE2990, GSE12093, GSE5327, GSE1456, GSE2034, GSE3494, GSE6532 and GSE9195). Based on the expression data reordered by PROB, we investigated which genes were upregulated or downregulated over progression by using a trend analysis technique. In addition, we confirmed that the expression levels of the above three genes, ASPM, CDCA8 and KIF2C, were significantly reduced following the knockdown or silencing of FOXM1 based on both microarray data in BT-20 breast cancer cells (GSE2222 [55]) (S10A–S10C Fig) and RNA-seq data in MCF-7 breast cancer cells (GSE58626 [56]) (S10D–S10F Fig). Gour et al. Considering technical variabilities in the sample-based transcriptomic data, it is important to have good robustness of the interaction coefficients in the GRN model with respect to the perturbation of the temporal progression. SGLT inhibition confers additional benefits in terms of HbA1c reduction, reduced glucose variability, small reduction in weight, and reduced total daily insulin doses without increasing the risk of hypoglycemia. C-peptide assays are commercially available, inexpensive, and widely accessible (Fig. While therapy with insulin is essential in all cases with undetectable C-peptide, patients diagnosed with LADA have, by definition, residual β-cell function and, in general, slow progression toward insulin dependency. Faced with a paucity of large-scale clinical trials in LADA, an expert panel reviewed data and delineated one therapeutic approach. Low latent inhibition Most people are able to ignore the constant stream of incoming stimuli, but this capability is reduced in those with low latent inhibition. We employed a set of scRNA-seq data of dendritic cells (DCs) for benchmarking since the gold standard in this situation is available. (e) A decrease in EMT score indicated a transition from epithelial to mesenchymal state during the progression of UC to SARC. However, it remains to be established whether insulin should be administered at an early stage of the clinical disease or whether it is the optimal therapy regardless of the stage of the disease process. A heatmap with hierarchical clustering (Fig 6A) showed that these 100 genes were clearly clustered into two groups: a descending group (purple) and an ascending group (blue). S can be viewed as a stage-weighted and locally scaled Gaussian kernel. Sarcomatoid urothelial bladder cancer (SARC) is a highly lethal variant of bladder cancer and has been reported to be evolved by the progression of the conventional urothelial carcinoma (UC) [39]. If (Xi(s), aij) and both satisfy the equations of progression-dependent dynamic model, i.e., The following report consists of three sections: 1) identifying subjects with LADA, 2) reviewing current therapeutic options, and 3) presenting the group’s proposal for management of LADA. (d) A Bayesian Lasso method is developed to infer the causal GRN based on the temporal data of gene expression. (c-e) ChIP-seq analysis of FOXM1 in the MCF-7 cell line with four biological replicates, showing that FOXM1 binds ASPM, CDCA8 and KIF2C. More information is available at https://www.diabetesjournals.org/content/license. One SGLT2i, dapagliflozin, has been recently approved by the European Medicines Agency for use in adults with T1D with BMI of at least 27 kg/m2 who failed to achieve adequate glycemic control despite optimal insulin therapy. Such a slowly evolving form of autoimmune diabetes, described as latent autoimmune diabetes of adults (LADA), accounts for 2–12% of all patients with adult-onset diabetes, though they show considerable variability according to their demographics and mode of ascertainment. In summary, we have developed a novel latent-temporal progression-based Bayesian Lasso method, PROB, to infer directed and signed gene networks from prevalent cross-sectional transcriptomic data. The kinetic features of the FOXM1-target interactions were formulated as a risk score to predict relapse for breast cancer patients in multiple independent cohorts. Autoantibodies against the IA-2(256–760) fragment were shown to be a reliable marker of LADA with a sensitivity and specificity of 40% and 97%, respectively (37). (f-h) ChIP-seq analysis of FOXM1 in the MDA-MB-231 cell line with two biological replicates, showing that FOXM1 binds ASPM, CDCA8 and KIF2C. Yes To this end, we propose that the latent temporal order of cancer progression status (i.e., latent-temporal progression) could be estimated from the cross-sectional data based on transcriptomic similarity between patient samples. https://doi.org/10.1371/journal.pcbi.1008379.g006. e1008379. Furthermore, here, we propose a GRN kinetic signature (S8 Text) based on FOXM1-targeted gene interactions to prognosticate relapse in breast cancer. (b) Gene set enrichment analysis for the descending genes (upper panel) and ascending genes (lower panel). In addition, we also included some GRN inference methods recently developed for scRNA-seq data into benchmarking analysis, since scRNA-seq data is also cross-sectional type. Importantly, considering the heterogeneity of LADA, for the clinician it is not possible to be certain that any given patient does or does not have LADA without estimating diabetes-associated autoantibodies. The cells were stimulated with LPS and sequenced at 1, 2, 4, and 6h after stimulation. (a) Simulated cross-sectional gene expression data. The clinical guidelines for the management of hyperglycemia in T2D do not take into account the diverse metabolic phenotypes of LADA. Mutual information (MI) measures non-linear dependency between genes and thus provides a natural generalization of the correlation. The UC-specific network and SARC-specific network were then constructed by extracting edges that were unique to UC network and SARC network respectively.