GEMap: A comprehensive gene essentiality map for drug discovery
1. Motivation
Gene dependency (synonymous with essentiality) quantifies the degree to which a gene is required for cellular proliferation and survival, exhibiting profound genetic context-dependence in malignant cells. Precisely quantifying these context-specific dependencies, high-throughput CRISPR-Cas9 essentiality screens have transformed precision oncology through systematic identification of tumor-specific vulnerabilities coupled with functional validation of therapeutic targets. However, critical limitations persist: CRISPR essentiality data remain fragmented across sources, and existing platforms lack intuitive mining tools. To address these challenges, we developed GEMap - an innovative web platform designed to consolidate human CRISPR-Cas9 essentiality screens and nearly 20,000 drug entries, reveal the gene essentiality landscape across thousands of molecularly annotated cancer models, and provide an interactive, customizable toolbox for researchers. GEMap can be used to (1) explore multimodal gene data (Dependency/Expression/CNV/Mutation) across cell lines and tissues; (2) identify potential therapeutic targets for patients with cancer classified by molecular features, such KRAS mutation; and (3) discover tailored treatments and drugs candidate targeting particular genes using large-scale drug perturbation data and drug physical targets. To the best of our knowledge, GEMAP represents the first integrated platform combining genome-wide CRISPR screening data with drug perturbation transcriptomic profiles and target interaction networks, establishing a novel systematic framework for therapeutic candidate discovery.
2. Overview
GEMap incorporates a total of 1,912 screens across 1,135 cell lines in 27 tissue types and includes four main modules: Essentiality Analysis, Essentiality2Target, Essentiality2Drug and DIY Tools.
2.1 Data
GEMap consolidates CRISPR screening data from Broad Achilles 23Q2, Sanger, Avana, GeCKO and supplements these with manually curated literature-derived datasets from PubMed. Drug data were sourced from LINCS 2020 and DGIdb datasets，including nearly 20,000 drug entries, enabling users to easily identify candidate drugs for genes of interest.
To ensure transparency and reproducibility, we provide a detailed description of each data source and its corresponding input options:
DepMap 23Q2 Data:
Source: DepMap Portal (https://depmap.org/portal/data_page/?tab=allData)
Dataset: DepMap Public 23Q2 Files – CRISPRGeneEffect.csv
Description: Genome-wide CRISPR-Cas9 knockout screens to assess gene essentiality across cancer cell lines.
Sanger CRISPR Data:
Source: DepMap Portal (https://depmap.org/portal/data_page/?tab=allData)
Dataset: Sanger CRISPR (Project Score, CERES) – gene_effect.csv
Description: CRISPR-Cas9 knockout data from the Sanger Institute, including CERES-corrected gene effect scores.
Avana Data:
Source: DepMap Portal (https://depmap.org/portal/data_page/?tab=allData)
Dataset: Ceres-gene-effect.csv
Description: Results of genome-wide CRISPR-Cas9 screens with the Avana sgRNA library in 341 cancer cell lines performed at the Broad Institute as part of the Cancer Dependency Map project.
GeCKO Data:
Source: DepMap Portal (https://depmap.org/portal/data_page/?tab=allData)
Dataset: gene_effect.csv
Description: 43 lines screened with the GeCKO CRISPR knockout library with gene effect scores inferred through the CERES algorithm.
PubMed Data:
Source: Curated from 62 peer-reviewed articles.
Description: Literature-derived data supporting gene-disease associations and therapeutic targets.
CCLE Data: Expression Data:
Source: Broad Institute CCLE (https://sites.broadinstitute.org/ccle)
Dataset: OmicsExpressionProteinCodingGenesTPMLogp1.csv
Description: Transcriptomic profiling of protein-coding genes across cancer cell lines.
CCLE Data: Mutation Data:
Source: Broad Institute CCLE (https://sites.broadinstitute.org/ccle)
Dataset: OmicsSomaticMutations.csv
Description: Somatic mutation profiles across cancer cell lines.
CCLE Data: Copy Number Data:
Source: Broad Institute CCLE (https://sites.broadinstitute.org/ccle) Dataset: OmicsCNGene.csv Description: Copy number variation data for genes across cancer cell lines.
DPP(Drug Perturbational Profiles) Data:
Source: LINCS 2020 (https://clue.io/releases/data-dashboard)
Dataset: level5_beta_trt_cp_n720216x12328.gctx
Description: Drug perturbation data, filtered for treatments with pert_itime = “24 h” and pert_idose = “10 uM”.
DTI(Drug Targets Interaction) Data:
Source: DGIdb (https://dgidb.org/)
Dataset: interactions.tsv
Description: We selected the latest data (2025/03/20). For drug target information, we mainly extracted and filtered drug-gene interactions by 10 interaction types, categorizing them as activating or inhibitory.
2.2 Analysis Workflows
GEMap streamlines data integration and reuse by offering not only basic browsing, retrieval, and download capabilities, but also a comprehensive suite of analysis workflows.
2.2.1 Essentiality Analysis consists of four parts:
User can explore the essentiality score, expression level, CNV level and mutation level of interesting gene by searching gene symbols, and view detailed data annotations of individual cell lines.
1. Dependency Analysis: User can view the dependence value of the gene of interest in different tissues and cell lines.
2. Expression Analysis: Users can view the expression value of the gene of interest in different tissues and cell lines.
3. CNV Analysis: Users can view the CNV value of the gene of interest in different tissues and cell lines, which can be amplification (copy number gain) or deletion (copy number loss) of the gene.
4. Mutation Analysis: Users can view the frequency and distribution of gene mutations through bubbles of different sizes and color gradients.
2.2.2 Essentiality2Target consists of two parts:
We used dependency-based therapeutic target prioritization pipeline and allows users to screen and visualize protentional therapeutic target in the context of gene copy number variations or mutations relevant to the user’s interests. If users aim to identify which cancer type with gene mutation/CNV exhibits greater sensitivity to this therapeutic target, users can navigate to the Tissue section.
1. Pancancer: User can identify protentional therapeutic target in the context of gene copy number variations or mutations relevant to the user’s interests in Pancancer mode.
2. Tissue: User can identify tissue-specific protentional therapeutic target in the context of gene copy number variations or mutations relevant to the user’s interests in certain cancer type.
2.2.3 Essentiality2Drug consists of two parts:
This module allows users to connect genes of interest to drugs, helping users to evaluate the most promising drug candidates based on gene perturbation effects and drug-target interactions.
1. DPP: Aims to identify drugs that regulate the target in specific tissues.
2. DTI: Aims to identify potential drugs that target specific genes.
2.2.4 DIY Tools consists of five parts:
DIY Tools can be categorized into Correlation and Visualize Your Way. The Correlation includes: Co-Dependency, Dependency/Expression, Dependency/CNV, Dependency/Mutation. It has been reported genes with correlated knockout fitness profiles tend to operate in the same biological process or pathway1 − 2. Additionally, the Visualize You Way module can assess both dependency and expression patterns across multiple genes.
1. Co-Dependency: Users can examines the correlation between two genes from the Dependency dataset.
2. Dependency-Expression: Users can examines the correlation between two genes, one from Dependency and the other from Expression.
3. Dependency-CNV: Users can examines the correlation between two genes, one from Dependency and the other from CNV.
4. Dependency-Mutation: Users can examines the impact of one gene's mutation on the dependency of another gene across various tissues.
5. Visualize You Way: Users can examine gene dependence and expression in different tissues for multiple genes.
3. Essentiality Analysis
The Essentiality Analysis includes four parts: Dependency, Expression, CNV, and Mutation.
3.1 Dependency
This section aims to allow user to explore the impact of a gene on cell growth or survival in different tissues and cell lines based on its gene dependency value. A lower Dependency value generally indicates that the gene is crucial for cell growth or survival. These data are valuable for understanding the role of genes across different tissues and their potential as therapeutic targets.
Follow these steps:
① Select the DataSet — DepMap 23Q2, Avana, GeCKO, Sanger or PubMed.
② Select a gene.
③ You can select either Pancancer or single cancer. If Single cancer is selected, step ⑥ will appear for choosing the specific cancer type.
④ You can select "Demo" for an example.
⑤ Click on “Quick Tip” to view the steps and content for this section.
If steps ①, ②, and ③ are selected, click "GO!" to view the following result:
For example, using CDK4, we plot boxplots for each cancer type, sorted by the median of Dependency values. A lower Dependency value typically indicates higher dependency of the gene on the cancer type. From Figure 1A, it is evident that CDK4 has higher dependency in the Bone tissue. Hovering over the boxplot provides interactive details, such as max, upper fence, Q3, median, Q1, lower fence, and min values (Figure 1A).
Next, we plot the Dependency values of CDK4 across all cell lines, allowing us to see which tissues and cell lines show higher gene dependency. This plot is also interactive, and you can display Tissue, Cell line, Essentiality Score (Figure 1B).
You can also click on "Plot result" and download the plot as a PDF.
If steps ①, ②, ③ and ⑥ are selected, click "GO!" to view the following result:
For CDK4, a boxplot for the Bone cancer type, highlighting significant sensitivity effects, is shown in Figure 2A. This plot shows the sensitivity of CDK4 in Bone tissue. Hovering over the plot reveals detailed statistics such as max, upper fence, Q3, median, Q1, lower fence, and min values (Figure 2A).
We then plot the Dependency values for CDK4 across all cell lines in the Bone tissue, allowing us to see which cell lines show the highest dependency. The plot is interactive, displaying Tissue, CellName, Essentiality Score (Figure 2B).
Click "Plot result" below to download the PDF version.
3.2 Expression
This section aims to reflects gene expression levels in different cancer types or celllines.
Follow these steps:
① DataSet — CCLE.
② Select a gene.
③ You can select either Pancancer or single cancer. If Single cancer is selected, step ⑥ will appear for choosing the specific cancer type.
④ You can select "Demo" for an example.
⑤ Click on “Quick Tip” to view the steps and content for this section.
If steps ② and ③ are selected, click "GO!" to view the following result:
For example, using CDK4, we plot boxplots for each cancer type, sorted by the median of Expression values. Higher Expression values typically indicate greater gene expression in the cancer type. Figure 3A shows that CDK4 has a higher expression in Soft Tissue cancer type. Hovering over the boxplot provides interactive details, such as the max, upper fence, Q3, median, Q1, lower fence, and min values (Figure 3A).
We then plot the Expression values of CDK4 across all cell lines, allowing us to see which tissues and cell lines show the highest expression levels. This plot is interactive, enabling you to display Tissue, Cell line, Expression Level (Figure 3B).
You can click "Plot result" and download the plot as a PDF.
If steps ②, ③ and ⑥ are selected, click "GO!" to view the following result:
For CDK4, we plot a boxplot for Soft Tissue cancer type, highlighting significant expression effects, as shown in Figure 4A. This plot provides an overview of CDK4 expression levels in Soft Tissue tissue. Hovering over the plot reveals detailed statistics, including the max, upper fence, Q3, median, Q1, lower fence, and min values (Figure 4A).
We then plot the Expression values for CDK4 across all cell lines in the Soft Tissue tissue, allowing us to see which cell lines show the highest expression levels. The plot is interactive, displaying Tissue, Cell line, Expression Level (Figure 4B).
Click "Plot result" below to download the PDF version.
3.3 CNV
This section aims to reflect the gene copy number variation (CNV) across tissues and celllines, including amplifications and deletions. These variations are key drivers of cancer initiation and progression.
The copy number was denoted gain if log2(CN ratio +1) was greater than 1.6 and a loss if less than 0.63.
Follow these steps:
① DataSet — CCLE.
② Select a gene.
③ You can select either Pancancer or single cancer. If Single cancer is selected, step ⑥ will appear for choosing the specific cancer type.
④ You can select "Demo" for an example.
⑤ Click on “Quick Tip” to view the steps and content for this section.
If steps ② and ③ are selected, click "GO!" to view the following result:
Using CDK4 as an example, we plot boxplots for each cancer type, sorted by the median Copy Number. This allows us to visually compare the variations in copy numbers across different cancers. Higher Copy Number values typically indicate an increase in gene expression for that cancer type. Figure 5A shows that CDK4 has a higher copy number in the Testis tissue. Hovering over the boxplot enables interaction, statistic results such as max, upper fence, Q3, median, Q1, lower fence, and min values (Figure 5A).
Next, we plot the Copy Number values of CDK4 across all cell lines, allowing us to see which tissues and cell lines have higher copy number levels. The plot is interactive, displaying Tissue, Cell line, CNV Level (Figure 5B).
You can click "Plot result" and download the plot as a PDF.
If steps ②, ③ and ⑥ are selected, click "GO!" to view the following result:
For CDK4, we plot a boxplot for the Testis, highlighting significant copy number effects, as shown in Figure 6A. The figure provides an overview of CDK4 copy number levels in the Testis cancer type. Hovering over the boxplot enables interaction, statistic results such as max, upper fence, Q3, median, Q1, lower fence, and min values (Figure 6A).
Next, we plot the Copy Number values of CDK4 across all cell lines in the Testis tissue, allowing us to see which cell lines have higher copy number levels. The plot is interactive, displaying Tissue, Cell line, CNV Level (Figure 6B).
You can click "Plot result" and download the plot as a PDF.
3.4 Mutation
This section aims to display gene mutations in various tissues using a bubble chart. The size of the bubbles and the color gradients indicate the frequency and distribution of gene mutations, further revealing the diversity of mutations across different tissue types and their potential biological significance. This visualization method helps quickly identify genes with higher mutation frequencies in specific tissues, providing valuable clues for subsequent functional analysis.
Follow these steps:
① DataSet — CCLE.
② Input the genelist to display (up to 20 genes).
③ In the Gene List Validation, genes that exist in the database are marked with ✅, and those that do not are marked with ⛔️.
④ You can select "Demo" for an example.
⑤ Click on “Quick Tip” to view the steps and content for this section.
If steps ② are selected, click "GO!" to view the following result:
The figure illustrates the mutation frequencies and distributions of 16 genes—ADA, ATXN7L1, ATXN7L2, ATXN7L3, KAT2A, SGF29, SUPT7L, TAF5L, TAF6L, SUPT20H, TAF10, TAF12, SUPT3H, TADA2B, TADA3 and USP22—across various tissues. It provides an interactive visualization where users can explore the gene, cancer type, mutation count, and mutation proportion (Figure 7).
To download the PDF version, click "Download as PDF."
We also provide the Table result, which shows the calculat results of gene sets. as shown in the figure below:
4. Essentiality2Target
The Essentiality2Target is divided into two sections: Pancancer and Tissue, including CNV and Mutation data.
Here, wa developed a pipeline for prioritizing therapeutic targets based on dependency, which allows users to filter and visualize therapeutic targets in the context of gene CNV or mutations relevant to the user's interests. Using the identified therapeutic targets, we evaluated Gene-altered malignancies for sensitivity to target inhibition. For each tissue，we required at least three cell lines in two groups (MUT vs WT) and ranked by difference and sensitivity. The following is the algorithm flow chart (Figure 8).
Pancancer:
Here we assume that users are interested in G1 and want to find new candidate therapeutic targets under G1 mutations:
Step1: The cell lines were categorized into two groups: the G1 mutant group (X cell lines) and the wild-type group (Y cell lines).
Step 2: Compute the difference in mean dependency scores and the corresponding p-value for each gene between the G1 mutant and wild-type groups. To assess statistical significance, we apply the Wilcoxon Rank Sum Test to compare the two independent samples (MUT vs. WT) and derive a p-value.
Step3: A negative Difference value indicates that the gene in mutant group exhibits stronger dependency compared to the WT group. The Pvalue is the difference between MUT and WT using the Wilcoxon test for two independent samples and the p value is extracted. The final potential therapeutic target score prioritized candidate targets based on the following formula.
Tissue:
If users aim to identify which cancer type with G1 mutation exhibits greater sensitivity to this therapeutic target, users can navigate to the Tissue section:
Step1: The cell lines were categorized into two groups: the G1 mutant group (X cell lines) and the wild-type group (Y cell lines).
Step 2: We computed two metrics: Sensitivity (defined as the mean dependency score of T1 in the G1 mutant group) and Difference (the delta in mean dependency scores of T1 between the G1 mutant and wild-type groups). Larger negative values correlate with increased target sensitivity in the tissue.
Note:
MUT: G1 gene mutation group WT: G1 wild type group
For CNV (gain/loss), the procedure is similar to that described above.
4.1 Pancancer
This section aims to enable the user to predict and prioritize therapeutic targets based on gene mutations or copy number variations of interest.
Follow these steps:
① Select a source——CNV(Gain), CNV(Loss) or Mutation.
② Select a gene.
③ Select target include——ALL, DPP or DTI.
DPP: Only targets exist in DPP.
DTI: Only targets exist in DTI.
④ You can Select "Demo" for an example.
⑤ Click on “Quick Tip” to view the steps and content for this section.
If steps ①, ② and ③ are selected, click "GO!" to view the following result:
We chose ALL as an example for analysis.
We ranked the results according to the formula, where smaller Score values indicate greater significance. The top 10 points with Score values are highlighted (Figure 9A). These top-left dots are likely to be genes that become more sensitive in the group of gene copy number variations or gene mutations, which may be critical for cancer cell survival.
We also show the sensitivity of the top ten genes before and after KRAS copy number variation (Figure 9B). A smaller Value after gene copy number variation or gene mutation indicates that the gene becomes more sensitive, making it a potential target for further research.
To download the PDF version, click "Download as PDF."
We also provide the Table result, which shows the calculat results of KRAS and other genes. The results are sorted by the Score values calculation, as shown in the figure below:
4.2 Tissue
This section aims to enable the user to identify tissues with optimal sensitivity and specificity for predicted candidate targets based on mutations or copy number variations in the gene of interest.
Follow these steps:
① Select data type——CNV(Gain), CNV(Loss) or Mutation.
② Select a gene.
③ Select a targetgene.
④ Select target include——ALL or DPP.
DPP: Only tissues exist in DPP.
⑤ You can Select "Demo" for an example.
⑥ Click on “Quick Tip” to view the steps and content for this section.
If steps ①, ②, ③ and ④ are selected, click "GO!" to view the following result:
We chose ALL as an example for analysis.
In the Essentiality2Target, we use KRAS and RAF1 as examples. For each tissue, we required at least three cell lines in two groups (CNV vs WT). This reveals the differential impact of KRAS and RAF1 across various tissues (Figure 10).
The x-axis represents Sensitivity, while the y-axis represents Difference. Lower values indicate better results. This suggests that Myeloid cancers harboring KRAS copy number variations exhibit greater sensitivity to RAF1.
To download the PDF version, click "Download as PDF."
5. Essentiality2Drug
The Essentiality2Drug is divided into two sections: DPP and DTI.
DPP: DPP characterizes genome-wide expression regulatory relationships between drugs and genes, determining whether a drug upregulates or downregulates specific genes and quantifying the magnitude of such effects. The DPP data added tissue type annotations for the cell lines, excluding those without tissue type information. As a result, we obtained a total of 10,472 drugs7 − 8.
DTI: DTI refers to the binding of a pharmaceutical compound to a specific biological target’s active site, thereby modulating the target’s functional activity. The DTI data mainly extracted and filtered drug-gene interactions by 10 interaction types, categorizing them as inhibition (inhibitor, blocker, antibody, negative modulator, cleavage, inverse agonist) or activation (agonist, activator, positive modulator, vaccine).
After identifying potential therapeutic targets in the previous sections, if a target is present in DPP, we can screen drugs that regulate it within the database. If the target is found in DTI, we can help users discover potential drugs that target specific genes.
5.1 DPP
Follow these steps:
① Select a gene.
② Select drug type——ALL Drugs, Approved Drugs.
③ Select a tissue.
④ You can Select "Demo" for an example.
⑤ Click on “Quick Tip” to view the steps and content for this section.
If steps ①, ② and ③ are selected, click "GO!" to view the following result:
We selected ALL Drugs as a representative case for analysis.
We use RAF1 as an example to show the top 15 drugs with significantly upregulated or downregulated targets. Values less than 0 indicate down-regulation of the gene, and values greater than 0 indicate up-regulation. As can be seen from the figure, AZD-5438 down-regulates RAF1 (Figure 11).
To download the PDF version, click "Download as PDF" .
We also provide the Table result, which presents the computed values for RAF1. The results are sorted by Z-score, as shown in the figure below.
5.2 DTI
Follow these steps:
① Select a gene.
② Select drug type——ALL Drugs, Approved Drugs.
③ You can Select "Demo" for an example.
④ Click on “Quick Tip” to view the steps and content for this section.
If steps ① and ② are selected, click "GO!" to view the following result:
We selected ALL Drugs as a representative case for analysis.
Using RAF1 as an example, we identify potential drugs targeting it, with the drug-target interaction network accessible to the user (Figure 12).
To download the PDF version, click "Download as PDF."
We also provide the Table result, which presents the Gene, Drug, and Directionality. The results are sorted by Value, as shown in the figure below.
6. DIY Tools
The DIY Tools consists of five sections: Co-Dependency, Dependency-Expression, Dependency-CNV, Dependency-Mutation and Visualize You Way.
Co-Dependency, Dependency-Expression, Dependency-CNV: This section aims to examine correlations between two genes, with one gene representing Dependency and the other representing Dependency,Expression or CNV.
Dependency-Mutation: This section aims to examine how the dependency of one gene changes across different tissues after an interest gene mutation.
Visualize You Way: This section enables you to explore the Dependency, Expression, and CNV of a genelist across one or more tissues.
6.1 Dependency- Dependency, Dependency-Expression, Dependency-CNV
It has been observed that genes exhibiting correlated knockout fitness profiles are often involved in similar biological processes or pathways4 − 6. In the Co-Dependency, Dependency-Expression, Dependency-CNV, and Dependency-Mutation. Users can explore the essentiality relationship between a query gene and a comparison gene. This module presents the essentiality score of the query gene relative to the comparison gene across several different scales. By analyzing these levels, uers can gain a more comprehensive understanding of the functional interactions and potential joint involvement of these genes in cellular processes.
Follow these steps:
① Select a gene——from Dependency.
② Select a gene——from Dependency, Expression, and CNV (ensure that the same gene is not selected twice in the Dependency section).
③ You can select either Pancancer or Single cancer. If Single cancer is selected, step ⑥ will appear for choosing the specific cancer type.
④ You can select "Demo" for an example.
⑤ Click on “Quick Tip” to view the steps and content for this section.
If steps ①, ②, and ③ are selected, click "GO!" to view the following result:
Using A1BG and A2M as examples, correlation plots are generated for both the overall data and for each individual cancer type, sorted by correlation from low to high. Because the sample sizes for Vulva/Vagina and Testis are too small, we have placed them at the end by force (Figure 13). This allows for a clear visualization of the correlation differences across cancer types. Negative correlation typically suggests that higher expression of the gene in cancer cells may indicate stronger dependency. The figure shows that the negative correlation between A1BG and A2M in Ampulla of Vater tissues is relatively strong.
To download the PDF version, click "Download as PDF."
If steps ①, ②, ③ and ⑥ are selected, click "GO!" to view the following result:
Using A1BG and A2M as examples, we generate a plot showing a significantly negative correlation, which provides a clear visualization of their correlation in Ampulla of Vater tissue (Figure 14). A strong negative correlation usually indicates that a higher expression of the gene in cancer cells may correspond to a stronger dependence on the gene.
To download the PDF version, click "Download as PDF."
6.2 Dependency-Mutation
Follow these steps:
① Select a gene——from Dependency.
② Select a gene——from Mutation.
③ You can select either Pancancer or Single cancer. If Single cancer is selected, step ⑥ will appear for choosing the specific cancer type.
④ You can Select "Demo" for an example.
⑤ Click on “Quick Tip” to view the steps and content for this section.
If steps ①, ②, and ③ are selected, click "GO!" to view the following result:
Using A1BG mutation and A2M dependency data across pancancer and individual cancer types as an example, the figure illustrates how A1BG mutation affects A2M dependency in various cancer types (Figure 15). The plot shows that some cancer types do not have mutation data, so we focus only on those cancer types with available mutation data. If A1BG mutation results in a decrease in A2M dependency, it suggests increased sensitivity in that cancer type, making the research meaningful. The plot indicates that A2M dependency decreases in Soft tissue following A1BG mutation.
To download the PDF version, click "Download as PDF."
If steps ①, ②, ③ and ⑥ are selected, click "GO!" to view the following result:
Using the example of A1BG mutation and A2M dependency data in Soft tissue cancer type, we can observe that A1BG mutation leads to a decrease in A2M dependency value in Soft tissue, indicating increased sensitivity of the Soft tissue cancer type, which makes the research meaningful (Figure 16).
To download the PDF version, click "Download as PDF."
6.3 Visualize You Way
Follow these steps:
① Select a data source — Dependency, Expression and CNV.
② Enter the genelist; genes present in the database will be marked with a ✅, and those not present will be marked with a ⛔️.
③ Select one or more tissues.
④ You can Select "Demo" for an example.
⑤ Click on “Quick Tip” to view the steps and content for this section.
If steps ①, ② and ③ are selected, click GO! to view the result as shown below:
We use ADA, ATXN7L1, ATXN7L2, ATXN7L3, KAT2A, SGF29, SUPT7L, TAF5L, and the tissues Bone, Prostate, and Liver as examples to visualize the distribution of Dependency (Figure 17).
To download the PDF version, click "Download as PDF."
7. Download
Download You Way: This section allows you to download files.
Follow these steps:
① Select a data type.
② Select a data source.
③ Select genes or tissues.
④ Select one or more items.
⑤ You can Select "Demo" for an example.
If steps ①, ②, ③ and ④ are selected, click GO! to view the result as shown below:
A demo example is
provided to display the available data. To download the full dataset,
click "Download table."
Reference
1 Wainberg, M. et al. A genome-wide atlas of co-essential modules assigns function to uncharacterized genes. 53, 638-649 (2021).
2 Novak, L. C., Chou, J., Colic, M., Bristow, C. A. & Hart, T. J. N. a. r. PICKLES v3: the updated database of pooled in vitro CRISPR knockout library essentiality screens. 51, D1117-D1121 (2023).
3 Jaiswal A, Gautam P, Pietilä EA, Timonen S, Nordström N, Akimov Y, Sipari N, Tanoli Z, Fleischer T, Lehti K, Wennerberg K, Aittokallio T. Multi-modal meta-analysis of cancer cell line omics profiles identifies ECHDC1 as a novel breast tumor suppressor. Mol Syst Biol. 2021 Mar;17(3):e9526.
4 Hakobyan A, Meyenberg M, Vardazaryan N, et al. Pan-cancer Analysis of the Interplay Between Mutational Signatures and Cellular SignalingJ. ISCIENCE, 2024,27.
5 Tsherniak A, Vazquez F, Montgomery P G, et al. Defining a Cancer Dependency MapJ. Cell, 2017,170(3):564-576.
6 Shao X, Lv N, Liao J, et al. Copy Number Variation is Highly Correlated with Differential Gene Expression: a Pan-Cancer StudyJ. BMC Medical Genetics, 2019,20.
7 Tong X, Qu N, Kong X, et al. Deep representation learning of chemical-induced transcriptional profile for phenotype-based drug discoveryJ. Nat Commun, 2024,15(1):5378.
8 Broad Institute. Expanded CMap LINCS Resource 2020. https://clue.io/data/CMap2020#LINCS2020 (Accessed 20 February 2022).