Boosting Clear Cell Renal Carcinoma-Specific Drug Discovery Using a Deep Learning Algorithm and Single-Cell Analysis.

Clear cell renal carcinoma (ccRCC), the most common subtype of renal cell carcinoma, has the high heterogeneity of a highly complex tumor microenvironment. Existing clinical intervention strategies, such as target therapy and immunotherapy, have failed to achieve good therapeutic effects. In this article, single-cell transcriptome sequencing (scRNA-seq) data from six patients downloaded from the GEO database were adopted to describe the tumor microenvironment (TME) of ccRCC, including its T cells, tumor-associated macrophages (TAMs), endothelial cells (ECs), and cancer-associated fibroblasts (CAFs). Based on the differential typing of the TME, we identified tumor cell-specific regulatory programs that are mediated by three key transcription factors (TFs), whilst the TF EPAS1/HIF-2α was identified via drug virtual screening through our analysis of ccRCC's protein structure. Then, a combined deep graph neural network and machine learning algorithm were used to select anti-ccRCC compounds from bioactive compound libraries, including the FDA-approved drug library, natural product library, and human endogenous metabolite compound library. Finally, five compounds were obtained, including two FDA-approved drugs (flufenamic acid and fludarabine), one endogenous metabolite, one immunology/inflammation-related compound, and one inhibitor of DNA methyltransferase (N4-methylcytidine, a cytosine nucleoside analogue that, like zebularine, has the mechanism of inhibiting DNA methyltransferase). Based on the tumor microenvironment characteristics of ccRCC, five ccRCC-specific compounds were identified, which would give direction of the clinical treatment for ccRCC patients.

Identifying local associations in biological time series: algorithms, statistical significance, and applications.

Local associations refer to spatial-temporal correlations that emerge from the biological realm, such as time-dependent gene co-expression or seasonal interactions between microbes. One can reveal the intricate dynamics and inherent interactions of biological systems by examining the biological time series data for these associations. To accomplish this goal, local similarity analysis algorithms and statistical methods that facilitate the local alignment of time series and assess the significance of the resulting alignments have been developed. Although these algorithms were initially devised for gene expression analysis from microarrays, they have been adapted and accelerated for multi-omics next generation sequencing datasets, achieving high scientific impact. In this review, we present an overview of the historical developments and recent advances for local similarity analysis algorithms, their statistical properties, and real applications in analyzing biological time series data. The benchmark data and analysis scripts used in this review are freely available at http://github.com/labxscut/lsareview.

TIMELESS promotes the proliferation and migration of lung adenocarcinoma cells by activating EGFR through AMPK and SPHK1 regulation.

BACKGROUND: Lung adenocarcinoma (LUAD) has high morbidity and is prone to recurrence. TIMELESS (TIM), which regulates circadian rhythms in Drosophila, is highly expressed in various tumors. Its role in LUAD has gained attention, but the detailed function and mechanism have not been clarified completely at present. METHODS: Tumor samples from patients with LUAD patient data from public databases were used to confirm the relationship of TIM expression with lung cancer. LUAD cell lines were used and siRNA of TIM was adopted to knock down TIM expression in LUAD cells, and further cell proliferation, migration and colony formation were analyzed. By using Western blot and qPCR, we detected the influence of TIM on epidermal growth factor receptor (EGFR), sphingosine kinase 1 (SPHK1) and AMP-activated protein kinase (AMPK). With proteomics analysis, we comprehensively inspected the different changed proteins influenced by TIM and did global bioinformatic analysis. RESULTS: We found that TIM expression was elevated in LUAD and that this high expression was positively correlated with more advanced tumor pathological stages and shorter overall and disease-free survival. TIM knockdown inhibited EGFR activation and also AKT/mTOR phosphorylation. We also clarified that TIM regulated the activation of SPHK1 in LUAD cells. And with SPHK1 siRNA to knock down the expression level of SPHK1, we found that EGFR activation were inhibited greatly too. Quantitative proteomics techniques combined with bioinformatics analysis clarified the global molecular mechanisms regulated by TIM in LUAD. The results of proteomics suggested that mitochondrial translation elongation and termination were altered, which were closely related to the process of mitochondrial oxidative phosphorylation. We further confirmed that TIM knockdown reduced ATP content and promoted AMPK activation in LUAD cells. CONCLUSIONS: Our study revealed that siTIM could inhibit EGFR activation through activating AMPK and inhibiting SPHK1 expression, as well as influencing mitochondrial function and altering the ATP level; TIM's high expression in LUAD is an important factor and a potential key target in LUAD.

Ksak: A high-throughput tool for alignment-free phylogenetics.

Phylogenetic tools are fundamental to the studies of evolutionary relationships. In this paper, we present Ksak, a novel high-throughput tool for alignment-free phylogenetic analysis. Ksak computes the pairwise distance matrix between molecular sequences, using seven widely accepted k-mer based distance measures. Based on the distance matrix, Ksak constructs the phylogenetic tree with standard algorithms. When benchmarked with a golden standard 16S rRNA dataset, Ksak was found to be the most accurate tool among all five tools compared and was 19% more accurate than ClustalW2, a high-accuracy multiple sequence aligner. Above all, Ksak was tens to hundreds of times faster than ClustalW2, which helps eliminate the computation limit currently encountered in large-scale multiple sequence alignment. Ksak is freely available at https://github.com/labxscut/ksak.

Regularized survival learning and cross-database analysis enabled identification of colorectal cancer prognosis-related immune genes.

Colon adenocarcinoma is the most common type of colorectal cancer. The prognosis of advanced colorectal cancer patients who received treatment is still very poor. Therefore, identifying new biomarkers for prognosis prediction has important significance for improving treatment strategies. However, the power of biomarker analyses was limited by the used sample size of individual database. In this study, we combined Genotype-Tissue Expression (GTEx) and The Cancer Genome Atlas (TCGA) databases to expand the number of healthy tissue samples. We screened differentially expressed genes between the GTEx healthy samples and TCGA tumor samples. Subsequently, we applied least absolute shrinkage and selection operator (LASSO) regression and multivariate Cox analysis to identify nine prognosis-related immune genes: ANGPTL4, IDO1, NOX1, CXCL3, LTB4R, IL1RL2, CD72, NOS2, and NUDT6. We computed the risk scores of samples based on the expression levels of these genes and divided patients into high- and low-risk groups according to this risk score. Survival analysis results showed a significant difference in survival rate between the two risk groups. The high-risk group had a significantly lower overall survival rate and poorer prognosis. We found the receiver operating characteristic based on the risk score was showed to accurately predict patients' prognosis. These prognosis-related immune genes may be potential biomarkers for colorectal cancer diagnosis and treatment. Our open-source code is freely available from GitHub at https://github.com/gutmicrobes/Prognosis-model.git.

Tumor tissue microorganisms are closely associated with tumor immune subtypes.

Studies have found that different immune subtypes are present in the same tumor. Different tumor subtypes have different tumor microenvironments (TME). This means that the efficacy of immunotherapy in actual applications will, therefore, have different results. Existing tumor immune subtype studies have mostly focused on immune cells, stromal cells, genes and molecules without considering the presence of microbes. Some studies have shown that microflora can strongly promote many gastrointestinal cancers. The microbiome has, therefore, become an important biomarker and regulatory factor of cancer progression and therapeutic responses. In addition, the presence of microflora can strongly regulate the host immune system, indirectly affecting tumor growth. Taken together, it is important to study the relationships that develop among tumor tissue microorganisms, tumor immune subtype, and the TME. In this study, correlations between microbial abundance, immune cell infiltration, immune gene expression and tumor immune subtype were studied. To accomplish this, tissue microorganisms and immune cell ratios with significant differences between the different cancers were obtained by comparing 203 gastric cancer and intestinal cancer samples. Two immune subtypes of intestinal samples were obtained by K-means clustering algorithm and tissue microorganisms, immune cell ratios and immune-related genes with significant differences between different immune subtypes were screened through Wilcoxon rank sum test. The results showed that Clostridioides difficile, Aspergillus fumigatus, Yarrowia lipolytica, and Fusarium pseudograminearum were all closely associated with the identified tumor immune subtypes. Our open-source software is freely available from GitHub at https://github.com/gutmicrobes/IMM-subtype.git.

DPADM: a novel algorithm for detecting drug-pathway associations based on high-throughput transcriptional response to compounds.

Pathway genes functionally participate in the same biological process. They typically act cooperatively, and none is considered dispensable. The dominant paradigm in drug discovery is the one-to-one strategy, which aims to find the most sensitive drug to act on an individual target. However, many complex diseases, such as cancer, are caused by dysfunction among multiple-gene pathways, not just one. Therefore, identifying pathway genes that are responsive to synthetic compounds in a global physiological environment may be more effective in drug discovery. The high redundancy of crosstalk between biological pathways, though, hints that the covariance matrix, which only connects genes with strong marginal correlations, may miss higher-level interactions, such as group interactions. We herein report the development of DPADM-a Drug-Pathway association Detection Model that infers pathways responsive to specific drugs. This model elucidates higher-level gene-gene interactions by evaluating the conditional dependencies between genes under different drug treatments. The advantage of the proposed method is demonstrated using simulation studies by comparing with another two methods. We applied this model to the Connectivity Map data set (CMap), and demonstrated that DPADM is able to identify many drug-pathway associations, such as mitoxantrone (MTX)- PI3K/AKT association, which targets the topological conditions of DNA transcription. Surprisingly, apart from identifying pathways corresponding to specific drugs, our methodology also revealed new drug-related pathways with functions similarly to those of seed genes.

CTGF Promotes the Osteoblast Differentiation of Human Periodontal Ligament Stem Cells by Positively Regulating BMP2/Smad Signal Transduction.

Objective: This work is aimed at revealing the role and the molecular mechanism of connective tissue growth factor 2 (CTGF) in the osteoblast differentiation of periodontal ligament stem cells (PDLSCs). Methods: The osteogenic differentiation of PDLSCs was induced by osteogenic induction medium (OM), and the expression level of osteogenic related proteins ALP, RUNX2, OCN, and CTGF was estimated using quantitative real-time polymerase chain reaction (qRT-PCR) and Western blotting analysis. We constructed cell lines with CTGF overexpression or knockdown to verify the role of CTGF in the osteoblast differentiation of PDLSCs. Alkaline phosphatase (ALP) staining was introduced to measure the osteoblasts activity, and alizarin red S (ARS) staining was employed to test matrix mineralization. The interaction between CTGF and bone morphogenetic protein-2 (BMP-2) was determined by endogenous coimmunoprecipitation (Co-IP). Results: The expression level of CTGF was increased during the osteogenic induction of PDLSCs. Additionally, CTGF overexpression effectively maintained the stemness and facilitated the osteoblast differentiation in PDLSCs, and CTGF knockdown exerted opposite effects. Moreover, at molecular mechanism, CTGF increased the activity of BMP-2/Smad signaling pathway. Conclusion: This investigation verified that CTGF promotes the osteoblast differentiation in PDLSCs at least partly by activating BMP-2/Smad cascade signal.

The application of Caprini Risk Assessment Model in evaluation of deep vein thrombosis for patients with end-stage osteoarthritis before arthroplasty.

BACKGROUND: Deep vein thrombosis (DVT) was a fatal complication of knee arthroplasty. We had neglected the risk factors of preoperative DVT although patients undergoing knee arthroplasty were at high risk for VTE. This study was to determine the risk factors for preoperative DVT and application of Caprini Risk Assessment Model (RAM) in patients with end-stage knee osteoarthritis (OA). METHODS: We retrospectively analyzed 1808 cases with end-stage knee OA undergoing primary knee arthroplasty from May 2015 to December 2020. Based on the results of ultrasonography in lower extremities, all patients were divided into non-DVT group and DVT group. Distribution of risk factors and risk levels were compared using χ2 test between two groups. Binary logistic regression analysis was used to determine the risk factors and relationship of risk levels and preoperative DVT. RESULTS: The incidence of preoperative DVT was 5.53% (n = 100). Distribution of the study population by risk level was low, 4.09%; moderate, 23.95%; high, 66.98%; and highest 4.98%. Female (P = 0.002), age (P = 0.012), swollen legs (P = 0.035) and history of blood clots (P < 0.001) was correlated with preoperative DVT. Difference among four risk levels was significant (P = 0.007). Patients with highest risk level had statistically significant association with preoperative DVT (P = 0.005, OR = 2.93, 95%CI [1.375-6.246]). CONCLUSION: The incidence of preoperative DVT was 5.53% in end-stage knee OA patients. The gender (female) and age were independent risk factors for preoperative DVT. The risk group classification by Caprini RAM was significantly associated with preoperative DVT. The usage of Caprini RAM before knee arthroplasty may be beneficial for prophylaxis of DVT.

Predicting CRISPR/Cas9 Repair Outcomes by Attention-Based Deep Learning Framework.

As a simple and programmable nuclease-based genome editing tool, the CRISPR/Cas9 system has been widely used in target-gene repair and gene-expression regulation. The DNA mutation generated by CRISPR/Cas9-mediated double-strand breaks determines its biological and phenotypic effects. Experiments have demonstrated that CRISPR/Cas9-generated cellular-repair outcomes depend on local sequence features. Therefore, the repair outcomes after DNA break can be predicted by sequences near the cleavage sites. However, existing prediction methods rely on manually constructed features or insufficiently detailed prediction labels. They cannot satisfy clinical-level-prediction accuracy, which limit the performance of these models to existing knowledge about CRISPR/Cas9 editing. We predict 557 repair labels of DNA, covering the vast majority of Cas9-generated mutational outcomes, and build a deep learning model called Apindel, to predict CRISPR/Cas9 editing outcomes. Apindel, automatically, trains the sequence features of DNA with the GloVe model, introduces location information through Positional Encoding (PE), and embeds the trained-word vector matrixes into a deep learning model, containing BiLSTM and the Attention mechanism. Apindel has better performance and more detailed prediction categories than the most advanced DNA-mutation-predicting models. It, also, reveals that nucleotides at different positions relative to the cleavage sites have different influences on CRISPR/Cas9 editing outcomes.

Joint Analysis of Microbial and Immune Cell Abundance in Liver Cancer Tissue Using a Gene Expression Profile Deconvolution Algorithm Combined With Foreign Read Remapping.

Recent transcriptomics and metagenomics studies showed that tissue-infiltrating immune cells and bacteria interact with cancer cells to shape oncogenesis. This interaction and its effects remain to be elucidated. However, it is technically difficult to co-quantify immune cells and bacteria in their respective microenvironments. To address this challenge, we herein report the development of a complete a bioinformatics pipeline, which accurately estimates the number of infiltrating immune cells using a novel Particle Swarming Optimized Support Vector Regression (PSO-SVR) algorithm, and the number of infiltrating bacterial using foreign read remapping and the GRAMMy algorithm. It also performs systematic differential abundance analyses between tumor-normal pairs. We applied the pipeline to a collection of paired liver cancer tumor and normal samples, and we identified bacteria and immune cell species that were significantly different between tissues in terms of health status. Our analysis showed that this dual model of microbial and immune cell abundance had a better differentiation (84%) between healthy and diseased tissue. Caldatribacterium sp., Acidaminococcaceae sp., Planctopirus sp., Desulfobulbaceae sp.,Nocardia farcinica as well as regulatory T cells (Tregs), resting mast cells, monocytes, M2 macrophases, neutrophils were identified as significantly different (Mann Whitney Test, FDR< 0.05). Our open-source software is freely available from GitHub at https://github.com/gutmicrobes/PSO-SVR.git.

An Entity Relationship Extraction Model Based on BERT-BLSTM-CRF for Food Safety Domain.

Dealing with food safety issues in time through online public opinion incidents can reduce the impact of incidents and protect human health effectively. Therefore, by the smart technology of extracting the entity relationship of public opinion events in the food field, the knowledge graph of the food safety field is constructed to discover the relationship between food safety issues. To solve the problem of multi-entity relationships in food safety incident sentences for few-shot learning, this paper adopts the pipeline-type extraction method. Entity relationship is extracted from Bidirectional Encoder Representation from Transformers (BERTs) joined Bidirectional Long Short-Term Memory (BLSTM), namely, the BERT-BLSTM network model. Based on the entity relationship types extracted from the BERT-BLSTM model and the introduction of Chinese character features, an entity pair extraction model based on the BERT-BLSTM-conditional random field (CRF) is established. In this paper, several common deep neural network models are compared with the BERT-BLSTM-CRF model with a food public opinion events dataset. Experimental results show that the precision of the entity relationship extraction model based on BERT-BLSTM-CRF is 3.29%∼23.25% higher than that of other models in the food public opinion events dataset, which verifies the validity and rationality of the model proposed in this paper.

The changes of the calf-vein deformation and femoral vein peak velocity during ankle pump exercise with or without graduated compression stockings.

OBJECTIVES: To analyze the changes of lower limb hemodynamics parameters before and after wearing graduated compression stockings (GCS) during ankle pump exercise in patients preparing for arthroplastic surgery. METHOD: The leg veins of 16 patients awaiting arthroplasty were analyzed using a Sonosite M-Turbo ultrasound system during ankle pump exercise with or without GCS. The age of them was 70 ± 7 years (mean ± SD) (range 56-82 years) and body mass index was 25.8 ± 3.0 kg/m2 (range 18.0-30.5 kg/m2). Measured data including the cross-sectional area (CSA), anteroposterior (AP) diameter and lateromedial (LM) diameter of the soleus vein (SV), posterior tibial vein (PTV) and great saphenous vein (GSV). Additionally, the peak velocities of femoral vein (FV) were also measured. RESULTS: GCS could significantly decrease the cross-sectional area of SV, PTV and GSV in supine position at rest and maximum ankle plantar flexion. But the compression effect of GCS to SV and GSV was not observed during maximum ankle dorsiflexion. It was found that GCS application reduced the peak flow velocity of the femoral vein from 61.85 cm/s (95% CI = 50.94-72.75 cm/s) to 38.01 cm/s (95% CI = 28.42-47.59 cm/s) (P < 0.001) during ankle plantar flexion and decreased the femoral vein in these patients from 80.65 cm/s (95% CI = 70.37-90.92 cm/s) to 51.15 cm/s (95% CI = 42.58-59.73 cm/s) (P < 0.001) during ankle dorsiflexion. But this effect was not significant in supine position at rest. CONCLUSIONS: GCS could significantly reduce the peak flow velocity of the femoral vein during ankle pump exercise in the patients preparing for arthroplastic surgery.

CNN-XG: A Hybrid Framework for sgRNA On-Target Prediction.

As the third generation gene editing technology, Crispr/Cas9 has a wide range of applications. The success of Crispr depends on the editing of the target gene via a functional complex of sgRNA and Cas9 proteins. Therefore, highly specific and high on-target cleavage efficiency sgRNA can make this process more accurate and efficient. Although there are already many sophisticated machine learning or deep learning models to predict the on-target cleavage efficiency of sgRNA, prediction accuracy remains to be improved. XGBoost is good at classification as the ensemble model could overcome the deficiency of a single classifier to classify, and we would like to improve the prediction efficiency for sgRNA on-target activity by introducing XGBoost into the model. We present a novel machine learning framework which combines a convolutional neural network (CNN) and XGBoost to predict sgRNA on-target knockout efficacy. Our framework, called CNN-XG, is mainly composed of two parts: a feature extractor CNN is used to automatically extract features from sequences and predictor XGBoost is applied to predict features extracted after convolution. Experiments on commonly used datasets show that CNN-XG performed significantly better than other existing frameworks in the predicted classification mode.

An Entity Relation Extraction Method for Few-Shot Learning on the Food Health and Safety Domain.

In recent years, entity relation extraction has been a critical technique to help people analyze complex structured text data. However, there is no advanced research in food health and safety to help people analyze the complex concepts between food and human health and their relationships. This paper proposes an entity relation extraction method FHER for the few-shot learning in the food health and safety domain. For few-shot learning in the food health and safety domain, we propose three methods that effectively improve the performance of entity relationship extraction. The three methods are applied to the self-built data sets FH and MHD. The experimental results show that the method can effectively extract domain-related entities and their relations in a small sample size environment.

A Sparse and Low-Rank Regression Model for Identifying the Relationships Between DNA Methylation and Gene Expression Levels in Gastric Cancer and the Prediction of Prognosis.

DNA methylation is an important regulator of gene expression that can influence tumor heterogeneity and shows weak and varying expression levels among different genes. Gastric cancer (GC) is a highly heterogeneous cancer of the digestive system with a high mortality rate worldwide. The heterogeneous subtypes of GC lead to different prognoses. In this study, we explored the relationships between DNA methylation and gene expression levels by introducing a sparse low-rank regression model based on a GC dataset with 375 tumor samples and 32 normal samples from The Cancer Genome Atlas database. Differences in the DNA methylation levels and sites were found to be associated with differences in the expressed genes related to GC development. Overall, 29 methylation-driven genes were found to be related to the GC subtypes, and in the prognostic model, we explored five prognoses related to the methylation sites. Finally, based on a low-rank matrix, seven subgroups were identified with different methylation statuses. These specific classifications based on DNA methylation levels may help to account for heterogeneity and aid in personalized treatments.

Cell Heterogeneity Analysis in Single-Cell RNA-seq Data Using Mixture Exponential Graph and Markov Random Field Model.

Advanced single-cell profiling technologies promote exploration of cell heterogeneity, and clustering of single-cell RNA (scRNA-seq) data enables discovery of coexpression genes and network relationships between genes. In particular, single-cell profiling of circulating tumor cells (CTCs) can provide unique insights into tumor heterogeneity (including in triple-negative breast cancer (TNBC)), while scRNA-seq leads to better understanding of subclonal architecture and biological function. Despite numerous reports suggesting a direct correlation between circulating tumor cells (CTCs) and poor clinical outcomes, few studies have provided a thorough heterogeneity characterization of CTCs. In addition, TNBC is a disease with not only intertumor but also intratumor heterogeneity and represents various biological distinct subgroups that may have relationships with immune functions that are not clearly established yet. In this article, we introduce a new scheme for detecting genotypic characterization of single-cell heterogeneities and apply it to CTC and TNBC single-cell RNA-seq data. First, we use an existing mixture exponential family graph model to partition the cell-cell network; then, with the Markov random field model, we obtain more flexible network rewiring. Finally, we find the cell heterogeneity and network relationships according to different high coexpression gene modules in different cell subsets. Our results demonstrate that this scheme provides a reasonable and effective way to model different cell clusters and different biological enrichment gene clusters. Thus, using different internal coexpression genes of different cell clusters, we can infer the differences in tumor composition and diversity.

Efficacy of perioperative intravenous dexmedetomidine administration for the prevention of postoperative sore throat: a meta-analysis.

OBJECTIVE: Postoperative sore throat (POST) is an undesirable intubation-related complication after surgery. Several studies have investigated the efficacy of perioperative intravenous dexmedetomidine administration for the prevention of POST, but the results have been inconsistent. We aimed to summarize all existing evidence and draw a more precise conclusion to guide future clinical work. METHODS: PubMed, Cochrane Library, EMBASE and China National Knowledge Infrastructure databases were comprehensively searched for all randomized controlled trials published before 1 February 2021 that investigated the efficacy of dexmedetomidine for the prevention of POST. RESULTS: Nine studies involving 400 patients were included in our meta-analysis. Compared with the control groups (i.e., saline and anesthetic drugs), perioperative intravenous use of dexmedetomidine significantly reduced the incidence of POST [risk ratio (RR): 0.56; 95% confidence interval (CI): 0.40-0.77; I2 = 0%) and coughing on the tube during extubation (RR: 0.58; 95% CI: 0.41-0.82; I2 = 0%). Additionally, patients in the dexmedetomidine group were more likely to develop bradycardia (RR: 2.46; 95% CI: 1.28-4.71; I2 = 0%) and hypotension (RR: 3.26; 95% CI: 1.14-9.33; I2 = 0%) during the administration of dexmedetomidine than those in the control group. CONCLUSION: Perioperative intravenous administration of dexmedetomidine has a positive effect on the prevention of POST.

Colorectal Cancer Prediction Based on Weighted Gene Co-Expression Network Analysis and Variational Auto-Encoder.

An effective feature extraction method is key to improving the accuracy of a prediction model. From the Gene Expression Omnibus (GEO) database, which includes 13,487 genes, we obtained microarray gene expression data for 238 samples from colorectal cancer (CRC) samples and normal samples. Twelve gene modules were obtained by weighted gene co-expression network analysis (WGCNA) on 173 samples. By calculating the Pearson correlation coefficient (PCC) between the characteristic genes of each module and colorectal cancer, we obtained a key module that was highly correlated with CRC. We screened hub genes from the key module by considering module membership, gene significance, and intramodular connectivity. We selected 10 hub genes as a type of feature for the classifier. We used the variational autoencoder (VAE) for 1159 genes with significantly different expressions and mapped the data into a 10-dimensional representation, as another type of feature for the cancer classifier. The two types of features were applied to the support vector machines (SVM) classifier for CRC. The accuracy was 0.9692 with an AUC of 0.9981. The result shows a high accuracy of the two-step feature extraction method, which includes obtaining hub genes by WGCNA and a 10-dimensional representation by variational autoencoder (VAE).

Calculation of immune cell proportion from batch tumor gene expression profile based on support vector regression.

In addition to tumor cells, a large number of immune cells are found in the tumor microenvironment (TME) of cancer patients. Tumor-infiltrating immune cells play an important role in tumor progression and patient outcome. We improved the relative proportion estimation algorithm of immune cells based on RNA-seq gene expression profiling and solved the multiple linear regression model by support vector regression ([Formula: see text]-SVR). These steps resulted in increased robustness of the algorithm and more accurate calculation of the relative proportion of different immune cells in cancer tissues. This method was applied to the analysis of infiltrating immune cells based on 41 pairs of colorectal cancer tissues and normal solid tissues. Specifically, we compared the relative fractions of six types of immune cells in colorectal cancer tissues to those found in normal solid tissue samples. We found that tumor tissues contained a higher proportion of CD8 T cells and neutrophils, while B cells and monocytes were relatively low. Our pipeline for calculating immune cell proportion using gene expression profile data can be freely accessed from GitHub at https://github.com/gutmicrobes/EICS.git.