References

Several external resources were used or referenced in the creation of the the CCGD.

Studies

A list of all publications referenced in the content of this database and a description of their relevant findings are below.

Click here for a full bibliography in csv format.

Name	Description	Data
Bard-Chapeau 2014-01	To model hepatocellular carcinoma, Bard-Chapeau et al. used Sleeping Beauty to mutagenize transgenic mice expressing toxic HBV surface antigen in their livers. HBV infection is the major cause of human HCC, so this SB screen was designed to reveal the mutations that cooperate with liver inflammation to drive HCC. Following mutagenesis, 250 tumors from 34 mice were harvested and sequenced. Genetically related tumors were then excluded, leaving 228 tumors for analysis. All CISs on chromosome 1 were excluded due to local transposon hopping. Both gene-centric (gCIS) and Gaussian kernel convolution (GKC) methods were used to identify CISs. The final CIS list incorporated into this database represents the union of CISs identifed using both gCIS and GKC techniques.	Data were gathered from Supplementary Table 1a and 1b. CIS coordinates were converted from NCBI m37/mm9 to GRCm38/mm10 using UCSC LiftOver. For CISs identified by both GKC and gCIS analyses, the gCIS coordinates and rankings were used. Genes were ranked using p-Value from table 1a for gCIS and GKC Approximate p-value adjusted across genome from table 1b for GKC. If multiple genes were listed for a single CIS, each gene was given the same ranking and genomic coordinates.
Been 2014-01	Been et al., performed a forward genetic screen in mice to identify genes that contribute to histiocytic sarcoma. The model used the T2/Onc2 transposon, a conditional transposase (Rosa26-LsL-SB11) and a myeloid-specific Cre allele (LyzM-Cre). They used the TAPDANCE algorithm to identify CISs from insertions in 92 tumors from 36 mice.	Data was taken from Table 1. Relative rank was determined using the #Unique Regions in CIS column.
Berquam Vrieze 2011-01	Berquam-Vrieze et al. investigated the cell-of-origin in a mouse model of T-cell acute lymphoblastic leukemia by conducting Sleeping Beauty screens at various time points within the T-cell lineage. One of three different Cre transgenes activated RosaSBase-LSL, driving T2/Onc2 transposon integration in either HSCs (Vav-iCre), immature thymocytes lacking CD4/CD8 expression (Lck-Cre), or late-stage CD4+/CD8+ thymocytes (CD4-Cre). Screen 01 corresponds to insertions in Vav-iCre cells. All animals reaching 20 weeks of age in this cohort were euthanized. Genomic DNA used for analysis was extracted at the time of necropsy from thymic lymphomas.	All data were gathered from Supplementary Table 3. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. Predicted effect information was included in Table S3.
Berquam Vrieze 2011-02	Berquam-Vrieze et al. investigated the cell of origin in a mouse model of T-cell acute lymphoblastic leukemia by conducting Sleeping Beauty screens at various time points within the T-cell lineage. One of three different Cre transgenes activated RosaSBase-LSL, driving T2/Onc2 transposon integration in either HSCs (Vav-iCre), immature thymocytes lacking CD4/CD8 expression (Lck-Cre), or late-stage CD4+/CD8+ thymocytes (CD4-Cre). Screen 02 corresponds to insertions in Lck-Cre cells. Genomic DNA used for analysis was extracted at the time of necropsy from thymic lymphomas.	All data were gathered from Supplementary Table 3. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. Predicted effect information was included in Table S3.
Berquam Vrieze 2011-03	Berquam-Vrieze et al. investigated the cell of origin in a mouse model of T-cell acute lymphoblastic leukemia by conducting Sleeping Beauty screens at various time points within the T-cell lineage. One of three different Cre transgenes activated RosaSBase-LSL, driving T2/Onc2 transposon integration in either HSCs (Vav-iCre), immature thymocytes lacking CD4/CD8 expression (Lck-Cre), or late-stage CD4+/CD8+ thymocytes (CD4-Cre). Screen 03 corresponds to insertions in CD4-Cre cells. Genomic DNA used for analysis was extracted at the time of necropsy from thymic lymphomas.	All data were gathered from Supplementary Table 3. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. Three separate screens were conducted using HSCs (Vav-iCre), immature thymocytes (Lck-Cre), or late-stage CD4/CD8 double-positive thymocytes (CD4-Cre). Screen 03 corresponds to insertions in CD4-Cre cells. Predicted effect information was included in Table S3.
Chen 2016-01	Chen et al., used an organotypic recellulartization model by depleting cells from fresh normal human colon tissue and then seeding in hTERT-immortalized, but otherwise normal, human colonic epithelial cells (hCECs) along with primary myofibroblasts and endoethelial cells from healthy subjects. The SB screen was performed by stably transducing hCECs with an shRNA to APC followed by transfection with a modified version of T2/Onc2 along with SB100x. hCECs were then seeded into the model and monitored for submucosal invasion. 21 invasive neoplasis from 15 recellularized colon matrices were analyzed for SB insertions via linker-mediated PCR followed by sequencing. Insertion sites with the deepest read counts (top 10%) were considered clonal and were considered to identify candidate cancer genes.	Data was extracted from supplemental Table 2. Mouse homologs and GRCm38/mm10 genomic coordinates were identified using MGI Batch Query. Relative rank was assigned based on number of invasive neoplasias with insertions (1 = B, more than 1 = A). Unlike the majority of screens in this database, this was done using human cells and candidate genes were not identified by an anlysis of CISs.
Chen 2017-01	Chen, et al., performed an SB screen to identify drivers of breast cancer. They performed the screen using both predisposed mice using a mutant beta-catenin under control of the keratin 5 promoter (K5-N57β-cat) and wild-type mice. The SB system consisted of Rosa26-LsL-SB11 and T2/Onc and was activated by K5-Cre. They analyzed transposon insertions in 34 mammary tumors from both predisposed and non-predisposed mice. CISs were identified by gaussian kernel convolution.	Data was extracted from Dataset S2. CISs identifying the same gene were combined, while CISs identifying two genes were duplicated. CIS coordinates were converted to GRCm38/mm10 using UCSC LiftOver. Official symbols were obtained using MGI Batch Query. Relative Rank was based on GKC Approximate p-value adjusted across genome.
Collier 2005-01	Collier et al. tested the ability of the Sleeping Beauty system to pinpoint genes promoting solid tumor formation in mice. The SB10 transposase, driven by the ubiquitous CAGGS promoter, was used to enable T2/Onc transposition in the soma of transgenic mice. Crossing doubly-transgenic Sleeping Beauty mice with Arf-/- mice resulted in tumor formation. Linker-mediated PCR and sequencing identified transposon insertion sites residing in the genomic DNA extracted from tumorigenic tissues, and Monte Carlo-based analysis exposed statistically common sites.	All data were gathered from Supplementary Table 1. Start and End coordinates were originally reported using NCBI May 2004 build. They were converted to GRCm38/mm10 using UCSC LiftOver. LOC277923, the originally-assigned CIS name, was listed in the CCGD to distinguish it from the other non-gene CIS at chromosome 15 with coordinates that would not convert on LiftOver. No predicted effect information was included in the data. Regarding tumor type, lymphomas, malignant meningiomas, myeloid leukaemias and a pulmonary adenocarcinoma were observed, but the data were not shown.
Collier 2009-01	By generating mice with different transposon/transposase combinations, Collier et al. assessed the efficacy of Sleeping Beauty mutagenesis to accurately pinpoint candidate cancer drivers. Crossing T2/onc low-copy number lines with Rosa26-SB11 enabled the authors to successfully drive tumorigenesis in mice while avoiding the confounding problems of embryonic lethality and genomic instability. CIS analysis was performed after cloning insertions from 59 of the resulting leukemias/lymphomas and subtracting insertions located on the donor concatomer chromosome.	All data were gathered from Table 2. Start and End coordinates were originally reported using NCBI36 build. They were converted to GRCm38/mm10 using UCSC LiftOver. No predicted effect information was included in the data. Tumors were variable (lymphomas, leukemia, sarcoma, glioma).
Dorr 2015-01	Dorr et al., performed a forward genetic screen in mice to identify genes that contribute to lung cancer on a Pten deficient background. The model used the T2/Onc2 transposon, a conditional transposase (Rosa26-LsL-SB11), a lung epithelial cell-specific Cre allele (Spc-Cre), and a homozygous conditional Pten allele (Ptenfl/fl). They used the TAPDANCE algorithm to identify CISs from insertions in 23 tumors from 13 mice.	Data was taken from Table S3 Candidate Lung Cancer Genes based on CIS. UCSC LiftOver was used to convert genome coordinates to GRCm38/mm10. Number of insertions was used to calculate relative rank. The three Magi1 CISs were combined into a single CIS.
Dupuy 2005-01	Dupuy et al. enhanced the Sleeping Beauty mutagenesis system by creating the T2/Onc2 transposon vector and generating an SB11 knock-in allele targeted to the Rosa26 locus in mice. Successful SB transposition resulted in embryonic lethality and cancerous tumor formation in mice, demonstrating the utility of this transposon screen in higher eukaryotes for cancer gene identification. Common insertion sites were determined using a Java program that simulated random transposon insertions throughout the mouse genome.	All data were gathered from Supplementary Table 4. Start and End coordinates were originally reported using NCBI May 2004 build. They were converted to GRCm38/mm10 using UCSC LiftOver. No tumor type information was included in the CCGD due to the heterogeneity of tumors often associated with the same CIS-containing gene. No predicted effect information was included in the data.
Friedel 2013-01	Friedel et al. conducted a Piggybac transposon mutagenesis screen in mice harboring the ubiquitously expressed Rosa26-PB-transposase. Each Piggybac transposon array contained 20 transposon copies. In the experimental cohort of eight mice, 11 macroscopic tumors formed between 59 and 85 weeks of age. No specific cancer type was studied here, but resulting varieties included squamous cell carcinomas of the skin, lung and intestinal tumors, and follicular lymphoma of the spleen. To avoid PCR-generated bias toward short fragments for sequencing, tumor DNA was subjected to acoustic shearing.	Data were gathered from Figure 4. The Start and End coordinates correspond to the start and end locations of the gene containing the insertions. The coordinates are reported here according to GRC38/mm10.
Genovesi 2013-01	The objective of this study was to identify genes cooperating with sonic hedgehog to drive medulloblsatoma. To accomplish this, Genovesi et al. used Sleeping Beauty tumorigenesis in Patched1 heterozygous mice. The SB transposase was driven by the ubiquitous Rosa26 promoter. 85 medulloblastomas were PCR amplified and sequenced to search for common insertion sites. 77 total CISs were found, mapping to 56 different protein coding genes.	Data were gathered from Table S1. The Start and End coordinates correspond to the start and end locations of the gene containing the insertions. These coordinates are reported here according to GRC38/mm10. Predicted genes lacking identified transposon insertion coordinates were excluded from the final CIS list (U4, U6, U7).
Giotopoulos 2016-01	Giotopoulos et al., performed a Sleeping Beauty forward genetic screen to discover mutations that cooperate with the BCR-ABL translocation in CML to cause progression to blast crisis. To generate CML, the authors used a tetracycline-inducible (tet-off) BCR-ABL fusion transgene with tTA expressed under control of the TAL1 (SCL) enhancer to initiate CML. Sleeping Beauty transposase expression was initiated using a Rosa26-floxed-SB allele crossed to an Mx1-Cre allele and then mice were treated with pIpC to induce Mx1-Cre expression and tetracycline was discontinued to turn on BCR-ABL. All alleles were combined with a concatamer (80x) of the GrOnc transposon containing the Graffi1.4 virus LTR. Four cohorts of mice were generated (wt, SB only, BCR-ABL only, and BCR-ABL + SB). CISs were identified using Gaussian Kernel Convolution from sequences from 52 BCR-ABL + SB leukemias and from 20 SB only leukemias. This study (01) contains the CIS list from BCR-ABL + SB leukemias.	CIS genes were extracted from Supplemental Table 2. CIS genomic coordinates were based on the peak location and bp width. Genomic coordinates were converted to mm10 (GRCm38) using UCSC LiftOver tool. Two X-chromosome CISs did not convert and the two CIS’s identifying Irf2 were combined. Relative Rank was determined using the reported adjusted p-value (genome).
Giotopoulos 2016-02	Giotopoulos et al., performed a Sleeping Beauty forward genetic screen to discover mutations that cooperate with the BCR-ABL translocation in CML to cause progression to blast crisis. To generate CML, the authors used a tetracycline-inducible (tet-off) BCR-ABL fusion transgene with tTA expressed under control of the TAL1 (SCL) enhancer to initiate CML. Sleeping Beauty transposase expression was initiated using a Rosa26-floxed-SB allele crossed to an Mx1-Cre allele and then mice were treated with pIpC to induce Mx1-Cre expression and tetracycline was discontinued to turn on BCR-ABL. All alleles were combined with a concatamer (80x) of the GrOnc transposon containing the Graffi1.4 virus LTR. Four cohorts of mice were generated (wt, SB only, BCR-ABL only, and BCR-ABL + SB). CISs were identified using Gaussian Kernel Convolution from sequences from 52 BCR-ABL + SB leukemias and from 20 SB only leukemias. This study (02) contains the CIS list from SB only leukemias.	CIS genes were extracted from Supplemental Table 3. CIS genomic coordinates were based on the peak location and bp width. Genomic coordinates were converted to mm10 (GRCm38) using UCSC LiftOver tool. Relative Rank was determined using the reported adjusted p-value (genome).
Guo 2016-01	Guo et al., performed insertional mutagenesis by nucleofection of SB11 or SB100 along with T2/Onc into a non-tumorigenic Pro B-Cell line (BA/F3) and tested for cytokine independent growth and ability to form tumors in nude mice. They analyzed insertion sites in 1,100 colonies that became cytokine (IL3) independent using both Monte Carlo simulations and gCIS, resulting in the identification of ~5,500 CISs.	Data was taken from Supplementary Dataset 1. Official symbols and GRCm38/mm10 genomic coordinates were generated using MGI Batch query http://www.informatics.jax.org/batch. Duplicate genes were merged. Relative Rank was calculated using the lowest p-value (either MC-CIS of gCIS). Predicted effect was called gain if the authors SIB score had a p-value < 0.05.
Heltemes-Harris 2016-01	Heltemes-Harris et al., conducted a transposon insertional mutagenesis screen in mice to identify candidate cancer genes that contribute to B-Cell acute lymphoblastic leukemia. Their mouse model included a weakly constitutively active Stat5b transgene, which by itself was not sufficient to induce B cell leukemia in the majority of mice. They crossed the mice to CD79a-Cre x Rosa26-LsL-SB11 x T2/Onc to induce active transposition in progenitor B cells. They analyzed 65 leukemias and identified CISs using TAPDANCE and gCIS.	Data was extracted from Table 1, Supp Table 1 and Supp Table 2 and CISs were combined. Genomic coordinates were taken from TAPDANCE for 12 CISs and from gCIS for the remaining CISs. Coordinates were updated to GrcM38/mm10 using UCSC LiftOver and official symbols were based on MGI batch query. Relative rank was based on number of tumors with insertion.
Keng 2009-01	Keng et al. activated Sleeping Beauty transposition strictly within the liver to screen for genes commonly mutated in hepatocellular carcinoma. The SB11 transposase was knocked into the Rosa26 locus, along with a floxed-stop cassette, and activated by the hepatocellular-specific Albumin-Cre transgene. The T2/Onc transposon concatemer was efficiently excised and reintegrated using this system. A dominant-negative p53 transgene under conditional Cre control was also included in separate mice to generate mutations cooperating with the often-mutated p53 in HCC. Genomic DNA was extracted from tumorigenic mice and subsequent linker-mediated PCR and pyrosequencing identified transposon insertion sites. Common insertion sites were determined by comparison to Monte Carlo simulations. Common insertion sites from genetically predisposed mice (with p53 transgene) and non-predisposed mice were combined into one list.	Data were gathered from Table 1 for CIS details, and Table 3 for predicted effect information. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. No predicted effect information was included in the data.
Keng 2013-01	Keng et al. used a hepatocyte-specific albumin promoter driving Cre recombinase transgene to activate Sleeping Beauty transposition specifically in the liver. To mirror the most common mutations observed in HCC, a conditional dominant negative Trp53 transgene was also included. Tumors initiated via transposon mutagenesis in triple transgenic and quadruple transgenic (Trp53 mutant) mice were isolated and sequenced from both males and females and mutational differences between the two sexes examined.	Data from both genders and genetic backgrounds were pooled into 83 CICs harboring a total of 114 genes. Information was gathered from Table S7. Gene coordinates were originally reported using NCBI37 build. They were converted to GRC38/mm10 using UCSC LiftOver. Genomic coordinates for the CIS encompassing Msl2 would not convert using LiftOver, therefore, the coordinates listed in the database correspond to the start and end locations of the Msl2 gene.
Kodama 2016-01	Kodama et al., performed two forward genetic screens to identify hepatocellular cancer drivers using an albumin-cre transgene to mobilize SB T2/Onc3 with or without a conditional transgene expressing the hepatitis B virus surface antigen. They also included their data from a similar previous screen (see Bard-Chapeau 2014-01) which used SB T2/Onc2. Their final dataset incorporated both the SB T2/Onc2 and SB T2/Onc3 datasets and with or without the hepatitis B virus surface antigent. They used gCIS to identify CISs.	Data was extracted from supplementary Table 4. Official gene symbols and GRC38 (mm10) coordinates were converted using MGI Batch query. Relative rank was calculated using the lowest p-value for each gene.
Kodama 2016-02	Kodama et al., performed a forward genetic screen for liver cancer using hepatoblast cell lines generated from Albumin-Cre x Rosa26-LsL-SB11 x T2/Onc transgenic mice. The cell lines were injected into the flanks of Nude mice and resulted in liver tumors with an EMT signature. Hepatoblast cell lines generated from wildtype mice did not form tumors. Authors analyzed insertion sites in 52 tumors using gCIS to identify 223 candidate cancer genes.	CIS genes were taken from Supplemental Table S2. Official gene symbols and genomic coordinates for GRCm38/mm10 were extracted using MGI Batch querty (http://www.informatics.jax.org/batch). Relative rank was caclulated using p-values.
Koso 2012-01	Koso et al. were looking for driver mutations that transformed neural stem cells (NSCs) into glioma-initiating cells. Mutagenesis was achieved by crossing mice homozygous for a Rosa26-LSL-SB11 linked to T2/Onc2 or T2/Onc3 with Nes-Cre transgenics. The NSC-rich subventricular zones of embryonic brains with various genotypes were dissected out between E17 and P1 and cultured with EGF. EGF was later removed and replaced with FBS to induce differentiation. After a 2-week culture period, cells were re-seeded every week. Once the total cell number reached 108, cells were considered immortalized. Immortal cells were then injected into SCID mice, and GKC analysis was used to identify common insertion sites present in the resultant tumors.	All data were gathered from Dataset S3. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. Gene-centric analysis was performed by the authors, so the Start and End coordinates correspond to the start and end locations of the gene containing the insertions. No predicted effect information was included in the data.
Koso 2012-02	Koso et al. were looking for driver mutations that transformed neural stem cells (NSCs) into glioma-initiating cells. Mutagenesis achieved by crossing mice homozygous for a Rosa26-LSL-SB11 linked to T2/Onc2 or T2/Onc3 with Nes-Cre transgenics. The NSC-rich subventricular zones of embryonic brains with various genotypes dissected out between E17 and P1 and cultured with EGF. EGF was later removed and replaced with FBS to induce differentiation. After a 2-week culture period, cells were re-seeded every week. Once the total cell number reached 108, cells were considered immortalized. GKC analysis identified common insertion sites from immortal cells.	All data were gathered from Dataset S4. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. Gene-centric analysis was performed by the authors, so the Start and End coordinates correspond to the start and end locations of the gene containing the insertions. No predicted effect information was included in the data.
Koudijis 2011-01	Koudijis et al. tested a new method for analyzing transposon mutagenesis insertion sites, termed shear-splink. The authors’ claim was that by shearing genomic DNA into random fragments, shear-splink eliminated the bias associated with PCR amplification of fragments generated using restriction enzymes. A panel of 127 Sleeping Beauty-induced lymphomas was used for the methods comparison.	All data were gathered from Supplementary Table 6. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. Multiple restriction enzymes were used for mutagenic analysis, but the data was averaged into a single CCGD screen, which identified CISs based on NlaIII and BfaI-based RE-splink and the new analytic method, shear-splink. No predicted effect information was included in the data.
Lastowska 2013-01	Lastowska et al. identified mutations that cooperate with a heterozygous loss of Patched (Ptch+/-) to form medulloblastoma. Mice heterozygous for Ptch were crossed to Rosa26-SB11 and T2/Onc (low copy) lines to create Ptch+/- x SB11+/- x T2/Onc+/- mice. Mice developed medulloblastomas and other malignancies. Medulloblastomas were analyzed for SB insertions, and CISs were determined using the Gaussian kernel convolution method.	The CIS list is from Table 1A. CIS boundaries were determined from the supplemental BED file 2051-5960-1-35-s12.bed. Original mm9 coordinates were converted to mm10 using UCSC LiftOver. Relative rank was based on GKC p-value from Table 1A.
Mann 2012-01	Mann et al. identified mutations cooperating with oncogenic Kras in a mouse model of pancreatic cancer using the Sleeping Beauty transposon system. An oncogenic LSL-KrasG12D allele expressed specifically in the pancreas by a Pdx1-Cre driver was used in combination with T2/Onc2 and T3/Onc3 transgenic lines, in which Cre recombinase induced transposition via the Rosa26-LSL-SB11 cassette. Tumors were extracted and sequenced at necropsy from variable cohorts of mice and common insertion sites determined.	All data were gathered from Supplementary Table 1. Peak locations were listed for each CIS. This singular location was used as both the Start and End value. Coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. Predicted mutagenic consequences of insertion, listed in Table S1, are categorized as activation and disruption which are coded in the CCGD as Gain or Loss, respectively.
Mann 2015-01	To discover genes that cooperate with BrafV600E in melanoma, authors generated transgenic mice with a conditional BrafV600E allele (BrafCA/+) activated by a Tyr-creERT2 allele that is melanocyte specific. These mice were crossed with the conditional SB transposon alleles (five transposon mice, 2 versions of T2/Onc2 and 1 version of T2/Onc3) and Rosa26-LsL-SBase. Mice were on mixed backgrounds. Transposition was initiated by topical application of 4-hydroxytamoxifen. CISs were based on 70 sequenced melanomas and were identified using multiple methods (Gaussian kernel convolution and gCIS).	Data was extracted from supplmental Table 2. Because multiple methods were used to detect CISs, CIS genomic coordinates were assigned based on gene location. Gene locations and official symbols were obtained using MGI Batch query (GRCm38). Predicted effect was based on “Predicted Mutagenic Consequence of SB insertion” generated by authors. Relative rank was calculated using number of tumors for each CIS method (GKC, gCIS0Kb and gCIS15Kb) and gene was assigned highest relative rank if identified using multiple methods.
Mann 2016-01	To model acute myeloid leukemia, authors generated Rosa26-LsL-SB11 x T2/Onc2 x Actb-Cre mice. These mice were crossed to Trp53 w/t, Trp53+/-, and Trp53(R172H), a dominant negative allele. Authors seequenced SB insertions in 168 mouse spleens from all three cohorts and identified common insertion sites using multiple statistical methods including gaussian kernel convolution and gene-CIS. All CISs were combined to generate a list of 466 candidate cancer genes.	Data was extracted from Supplementary Table 1. Because CIS’s were determined by multiple methods, the CIS location was set to the gene start and end based on GRCm38.p6. Relative rank was based on author’s call of Trunk driver (A), Genome Significant (B), with all other CISs ranked as (C).
March 2011-01	March et al. located potential colorectal cancer driver genes using Sleeping Beauty mutagenesis in mice also carrying either germline or somatic Apc mutations in a C57BL/6J background. Cre induction in AhCre;Apcfloxed/wt;Rosa26Lox66SBLox71;T2/Onc mice simultaneously removed Apc and activated SB transposase, resulting in tumor formation largely in the small intestine. Tumor excision and subsequent sequencing revealed insertions sites analyzed using both GKC and MC methods.	All data were gathered from Supplementary Table 2C. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. Insertions from Apcmin and Apcfloxed models were combined for statistical calculations. CIS peaks were assigned to associated genes if the kernel peak fell within the coding region of a gene or was within 150 kb of a gene. No predicted effect information was included in the data. If a gene appeared in more than one CIS in the database, the CIS listed in the CCGD corresponds to that with the highest peak height value.
Montero-Conde 2017-01	To model HRas-driven thyroid cancer, Montero-Conde et al., generated transgenic mice with the following alleles: FR-HRasG12V (contains a w/t floxed HRAS allele followed by the HrasG12V allele, which will be expressed after the w/t allele is removed by CRE) x TPO-Cre (throid peroxidase promoter driven Cre) x Rosa-LsL-SB x T6113 (containing 358 T2/Onc2 transposons on Chr 1 from Dupuy, et al., 2009). 23 thyroid nodules from 19 mice were subjected to linker-mediated PCR and CISs were called using gCIS with the additional requirement that the insertion be in at least 3 mice.	Data was taken from Figure 1D listing the 45 gCIS genes and the number of tumors in which there were insertions. Genomic coordinates were obtained using MGI’s batch query. Relative rank was based on number of tumors per gCIS.
Moriarity 2015-01	Moriarity et al., used a bone-specific Cre (Osx-Cre) to activate SB11 transposase in bone precursors (Rosa26-LsL-SB11) in the presence of T2/Onc. Approximately 24% of mice developed osteosarcomas. Tumors from 26 of these mice were sequenced. They performed a second screen using mice with a conditional dominant negative Trp53 allele (Trp53-LsL-R270H) and accelerated tumor penetrance from 59% (Trp53 only) to 75% (Trp53 + SB mutagenesis). Tumors from 96 of these animals were sequenced. Finally, 19 mice with metastatic tumors were also sequenced. CISs were determined using TAPDANCE and a Gene-centric analysis. This study (01) is the SB x OSX-Cre screen.	Data was taken from Supplemental Table 2. The CIS list is a combination of TAPDANCE and gene centric analysis, so CIS coordinates are listed as the gene coordinates instead of the actual CIS coordinates. Relative Rank was determined using % of Tumors with CIS (If the percentage was different when comparing TAPDANCE to gene centric, the higher percentage was used).
Moriarity 2015-02	Moriarity et al., used a bone-specific Cre (Osx-Cre) to activate SB11 transposase in bone precursors (Rosa26-LsL-SB11) in the presence of T2/Onc. Approximately 24% of mice developed osteosarcomas. Tumors from 26 of these mice were sequenced. They performed a second screen using mice with a conditional dominant negative Trp53 allele (Trp53-LsL-R270H) and accelerated tumor penetrance from 59% (Trp53 only) to 75% (Trp53 + SB mutagenesis). Tumors from 96 of these animals were sequenced. Finally, 19 mice with metastatic tumors were also sequenced. CISs were determined using TAPDANCE and a gene-centric analysis. This study (02) is the Trp53mut x SB x OSX-Cre screen.	Data was taken from Supplemental Table 1. The CIS list is a combination of TAPDANCE and gene centric analysis, so CIS coordinates are listed as the gene coordinates instead of the actual CIS coordinates. Relative Rank was determined using % of Tumors with CIS (If the percentage was different when comparing TAPDANCE to gene centric, the higher percentage was used).
Moriarity 2015-03	Moriarity et al., used a bone-specific Cre (Osx-Cre) to activate SB11 transposase in bone precursors (Rosa26-LsL-SB11) in the presence of T2/Onc. Approximately 24% of mice developed osteosarcomas. Tumors from 26 of these mice were sequenced. They performed a second screen using mice with a conditional dominant negative Trp53 allele (Trp53-LsL-R270H) and accelerated tumor penetrance from 59% (Trp53 only) to 75% (Trp53 + SB mutagenesis). Tumors from 96 of these animals were sequenced. Finally, metastases from 19 mice were also sequenced. CISs were determined using TAPDANCE and a gene-centric analysis. This study (03) is the metastatic tumor screen.	Data was taken from Supplemental Table 12. The CIS list is a combination of TAPDANCE and gene centric analysis, so CIS coordinates are listed as the gene coordinates instead of the actual CIS coordinates. Relative Rank was determined using % of Tumors with CIS (If the percentage was different when comparing TAPDANCE to gene centric, the higher percentage was used).
Morris 2016-01	Morris et al., performed a forward genetic screen in mice to identify drivers of GI tract cancer in a wild-type and TGFBR mutant background. They used Rosa26-LsL-SB11 mice crossed to two strains of T2/Onc2 mice (strain 6070 and 6113) crossed to Villin-Cre. The same screen was repeated after crossing to Tgfbr2fl/fl mice. 130 tumors from 38 mice were analyzed in the wild-type screen and 130 tumors from 40 mice were analyzed for the Tgfbr knockout screen. CISs were called using TAPDANCE and gCIS. All CIS lists were combined for entry into the CCGD database.	Data was taken from Supplemental data file 2, from the worksheets entitled TAP_all_tumors, gCIS_mutant, and gCIS_wild-type. Duplicates were merged and GRCm38/mm10 coordinates were updated using UCSC LiftOver (for Monte Carlo CIS) and MGI Batch Query (for gCIS). Gene was based on Nearest Gene by BED Tool column.
Ni 2013-01	Ni et al. used low-copy Piggybac transposons to screen for melanoma candidate genes while avoiding insertional hypermutation. The group used a conditional Braf knock-in mouse model, BrafCA. Upon addition of Cre recombinase, the functionally-wildtype BrafCA is converted to BrafV600E, an important initiating mutation in melanoma. Tamoxifen addition converted BrafCA specifically in the melanocytes. PB transposase initiated random transposition, with a maximum of seven mutator transposons per cell. Linker-mediated PCR and Illumina sequencing identified insertion sites within resulting tumors.	Data were gathered from Table 1. The Start and End coordinates correspond to the start and end locations of the gene containing the insertions. These coordinates are reported here according to GRC38/mm10.
O’Donnell 2012-01	O’Donnell et al. used Sleeping Beauty transposon mutagenesis to identify hepatocellular carcinoma-causing genes cooperating with oncogenic Myc. The ubiquitously-expressed Rosa26 promoter drove expression of the SB11 transposase in mice also expressing a tet-repressible Myc transgene, the T2/Onc transposon, and the LAPtTA tet-transactivator specific to the liver. At six weeks doxycycline removal initiated the system and palpation monitoring every two to three days continued until all mice were euthanized by 16 weeks of age. Ligation-mediated PCR and sequencing pinpointed transposon integration sites within the genomic DNA of extracted tumorigenic tissue. Monte Carlo analysis was used to define statistically significant sites of common insertion.	All data were gathered from Table 1. Gene-centric analysis identified CIS regions. The Start and End coordinates correspond to the start and end locations of the gene containing the insertions according to NCBI38 build. No coordinates were listed for Non-Gene CIS regions, so the CIS name corresponds to the original name given by the authors. No predicted effect information was included in the data. Number of independent (non-redundant) insertions was specified regarding relative strength.
Perez Mancera 2012-01	Perez-Mancera et al. conducted a Sleeping Beauty transposon mutagenesis screen in a mouse model of pancreatic ductal preneoplasia to pinpoint genes cooperating with mutant Kras to drive adenocarcinoma. TL1 embryonic stem cells were electroporated with pRosa26-LSL-SA-SB13-BGHpolyA constructs to knock in transposase expression. Compound mutant KrasLSL-G12D;Pdx1-cre;T2/Onc;Rosa26-LSL-SB13 mice developed pancreatic neoplasms for examination. Tumors and metastases from 103 mice at necropsy were scrutinized for common insertions using both GKC and MC analysis.	All data were gathered from Supplementary Table 3. MC and GKC analyses treated as two separate screens. Screen 01 corresponds to the GKC analysis. Multi-gene CIS regions were separated into two separate CIS genes if the listed peak height was located within both genes. If not, one gene was designated as an other gene. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. No predicted effect information was included in the data. If single genes were identified multiple times using the same analytic method (GKC or MC), the gene listing with the highest number of insertions was included in the table.
Perez Mancera 2012-02	Perez-Mancera et al. conducted a Sleeping Beauty transposon mutagenesis screen in a mouse model of pancreatic ductal preneoplasia to pinpoint genes cooperating with mutant Kras to drive adenocarcinoma. TL1 embryonic stem cells were electroporated with pRosa26-LSL-SA-SB13-BGHpolyA constructs to knock in transposase expression. Compound mutant KrasLSL-G12D;Pdx1-cre;T2/Onc;Rosa26-LSL-SB13 mice developed pancreatic neoplasms for examination. Tumors and metastases from 103 mice at necropsy were scrutinized for common insertions using both GKC and MC analysis.	All data were gathered from Supplementary Table 3. MC and GKC analyses treated as two separate screens. Screen 02 corresponds to the MC analysis. Multi-gene CIS regions were separated into two separate CIS genes if the listed peak height was located within both genes. If not, one gene was designated as an other gene. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. No predicted effect information was included in the data. If single genes were identified multiple times using the same analytic method (GKC or MC), the gene listing with the highest number of insertions was included in the table.
Perna 2015-01	Perna et al., performed a forward genetic screen to find genes that cooperate with a Braf mutation (BrafV618E) to cause melanoma. A second part of the screen involved treating the mice with a Braf inhibitor (PLX4720, a precursor of Vemurafenib) and then sequencing tumors that became resistant to the inhibitor. The model used a conditional Braf allele (LsL-BrafV618E), the T2/Onc transposon, a conditional transposase (Rosa26-LsL-SB13) and a skin-specific Cre (Tyr-CreERT2) activated by topical application of 4-OHT. This study (2015-01) includes the CIS genes from the first part of this study (no inhibitor).	Data was taken from the Primary Melanomas CIS list in the supplemental Excel spreadsheet. CIS Peak location was converted from NCBI37/mm9 to GRC38/mm10 using UCSC LiftOver tool. Relative rank was calculated using the P-value adjusted across genome. Gene names were changed to official RefSeq symbols.
Perna 2015-02	Perna et al., performed a forward genetic screen to find genes that cooperate with a Braf mutation (BrafV618E) to cause melanoma. A second part of the screen involved treating the mice with a Braf inhibitor (PLX4720, a precursor of Vemurafenib) and then sequencing tumors that became resistant to the inhibitor. The model used a conditional Braf allele (LsL-BrafV618E), the T2/Onc transposon, a conditional transposase (Rosa26-LsL-SB13) and a skin-specific Cre (Tyr-CreERT2) activated by topical application of 4-OHT. This study (2015-02) includes the CIS genes from the second part of this study (inhibitor resistant tumors).	Data for this study is from the Resistant Melanomas CIS list in the supplemental Excel spreadsheet. CIS Peak location was converted from NCBI37/mm9 to GRC38/mm10 using UCSC LiftOver tool. Relative rank was calculated using the P-value adjusted across genome. Gene names were changed to official RefSeq symbols.
Quintana 2012-01	Quintana et al. performed a Sleeping Beauty screen in mouse skin to identify the genes implicated in nonmelanoma skin cancer etiology. The bovine keratin K5 promoter was used to drive SB11 transposase expression, which activated T2/Onc2 transposition specifically in the epidermal stem cells. Two mouse lines harboring the T2/Onc2 transposon in different chromosomal locations were used in the study. High-throughput sequencing of the genomic DNA obtained from the resulting tumors revealed common insertion sites.	All data were gathered from Supplementary Table S2. The number of tumors corresponds to relative strength used in CCGD. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. Gene-centric analysis identified CIS regions. The Start and End coordinates correspond to the start and end locations of the gene containing the insertions. No predicted effect information was included in the data.
Rad 2010-01	Rad, et al., used a PiggyBac system to generate tumors in mice using various different transposons. They generated CIS data from 63 blood cancers from 3 transposon lines (ATP2-S1, ATP2-S2 & ATP2-H32). The transposons in these strains contained an MSCV LTR promoter. The tumors generated were T cell lymphomas (21%), myeloid leukemias (35%) and unclassified (remainder). CISs were identified using Gaussian Kernel convolution.	CIS data was taken from Table S6. Coordinates were changed from mm9 to mm10 using UCSC LiftOver. Gene names were manually changed to official NCBI symbols. Ranking was done using the p values in Table S6. Because many of these were 0, the rank of A and B was equivalent, so all genes in this category were ranked as A. The other external genes were included and listed as Not Ranked. Other external genes with ambiguous official NCBI symbols were excluded. CISs identifying the same gene were collapsed into a single CIS.
Rad 2015-01	Rad et al., used the PiggyBac system to identify cancer genes that cooperate with KrasG12D to form pancreatic cancer. They used a Rosa-LsL-PB transposase, the ATP1 transposon, and PDX1-Cre to activate the system. They identified CISs in 49 tumors using acoustic shearing, NGS, and analysis using either TAPDANCE or Gaussian kernel convolution. Only the TAPDANCE CIS findings are reported here because the article used these CISs for their analyses.	Data was extracted from Table S1, reporting 176 CISs defined by the TAPDANCE algorithm. CIS coordinates were converted from mm9 to mm10 using UCSC LiftOver. The authors did not specify a candidate gene, but instead listed all genes within the CIS. The CIS list was expanded to include all genes. Gene symbols were manually converted to official NCBI symbols. Gene symbols with no, or ambiguous, official NCBI symbols were discarded. CISs identifying the same gene were combined into a single CIS.
Rahrmann 2009-01	Rahrmann et al. investigated the candidate cancer genes driving prostate cancer using a Sleeping Beauty screen in mice. Utilizing Rosa26-SB11;T2/onc transgenic mice in both wildtype and Arf-/- backgrounds, the authors identified proliferation clusters in the prostatic epithelium. These clusters were laser captured and analyzed using linker-mediated PCR followed by sequencing.	All data were gathered from Supplementary Table 1. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. Gene-centric analysis was performed by the authors, so the Start and End coordinates correspond to the start and end locations of the gene containing the insertions. No predicted effect information was included in the data.
Rahrmann 2013-01	Rahrmann et al. used the Sleeping Beauty system to investigate the drivers of malignant peripheral nerve sheath tumors (MPNSTs). Transposition induction specifically in the Schwann cells was achieved using Cnp-cre and the T2/Onc concatemer in mice with loss of Trp53 function and/or EGFR overexpression. Linker-mediated PCR was performed on genomic DNA extracted from neurofibromas and grade 3 PNSTs. Statistically significant common insertion sites were determined using both gene-centric analysis and TAPDANCE.	Data were divided into four separate screens. Information for screen 01 was gathered from Table S2. Screen 01 corresponds to neurofibroma CISs determined using TAPDANCE. Gene coordinates were originally reported using NCBI36 build. They were converted to GRC38/mm10 using UCSC LiftOver. The Start and End coordinates correspond to the start and end locations of the gene containing the insertions. No predicted effect information was given for the neurofibroma CISs.
Rahrmann 2013-02	Rahrmann et al. used the Sleeping Beauty system to investigate the drivers of malignant peripheral nerve sheath tumors (MPNSTs). Transposition induction specifically in the Schwann cells was achieved using Cnp-cre and the T2/Onc concatemer in mice with loss of Trp53 function and/or EGFR overexpression. Linker-mediated PCR was performed on genomic DNA extracted from neurofibromas and grade 3 PNSTs. Statistically significant common insertion sites were determined using both gene-centric analysis and TAPDANCE.	Data were divided into four separate screens. Information for screen 02 was gathered from Table S2. Screen 02 corresponds to neurofibroma CISs determined using gene-centric analysis. Gene coordinates were originally reported using NCBI36 build. They were converted to GRC38/mm10 using UCSC LiftOver. The Start and End coordinates correspond to the start and end locations of the gene containing the insertions. No predicted effect information was given for the neurofibroma CISs.
Rahrmann 2013-03	Rahrmann et al. used the Sleeping Beauty system to investigate the drivers of malignant peripheral nerve sheath tumors (MPNSTs). Transposition induction specifically in the Schwann cells was achieved using Cnp-cre and the T2/Onc concatemer in mice with loss of Trp53 function and/or EGFR overexpression. Linker-mediated PCR was performed on genomic DNA extracted from neurofibromas and grade 3 PNSTs. Statistically significant common insertion sites were determined using both gene-centric analysis and TAPDANCE.	Data were divided into four separate screens. Information for screen 03 was gathered from Table 1. Screen 03 corresponds to MNST CISs determined using TAPDANCE. Gene coordinates were originally reported using NCBI36 build. They were converted to GRC38/mm10 using UCSC LiftOver. The Start and End coordinates correspond to the start and end locations of the gene containing the insertions. Predicted information was provided in Table 1 for the MNST CISs.
Rahrmann 2013-04	Rahrmann et al. used the Sleeping Beauty system to investigate the drivers of malignant peripheral nerve sheath tumors (MPNSTs). Transposition induction specifically in the Schwann cells was achieved using Cnp-cre and the T2/Onc concatemer in mice with loss of Trp53 function and/or EGFR overexpression. Linker-mediated PCR was performed on genomic DNA extracted from neurofibromas and grade 3 PNSTs. Statistically significant common insertion sites were determined using both gene-centric analysis and TAPDANCE.	Data were divided into four separate screens. Information for screen 04 was gathered from Table 1. Screen 04 corresponds to MNST CISs determined using gene-centric analysis. Gene coordinates were originally reported using NCBI36 build. They were converted to GRC38/mm10 using UCSC LiftOver. The Start and End coordinates correspond to the start and end locations of the gene containing the insertions. Predicted information was provided in Table 1 for the MNST CISs.
Rangel 2016-01	Rangel et al., performed an SB screen to identify drivers of triple negative breast cancer. They used conditional Pten knockout mice (Ptentm1Hwu/J) with the K5-Cre transgene along with SB11 (Rosa26-LsL-SB11) and two strains of transposon positive mice (T2/Onc2 6113, T2/Onc3 12740). Mice developed adenocarcinomas, adenomyoepitheliomas, and adenosquamous carcinomas. Authors analyzed insertions in 34 tumors (18 from T2/Onc2 and 15 from T2/Onc3) using Gaussian Kernel Convolution and gCIS and combined into a single list.	Data was extracted from Supplemental Table 2. Official symbols and genomic coordinates for GrcM38/mm10 were extracted from MGI Batch Query. Genomic coordinates for GKC calls were based on CIS Peak location +/- half of CIS width. Coordinates were updated to GrcM38/mm10 using UCSC LiftOver. Relative Rank was based on number of tumors with insertions, using GKC values if possible, otherwise using gCIS values.
Starr 2009-01	A Sleeping Beauty mutagenesis screen was performed on a wild-type mixed strain background using the villin promoter to activate transposition in the intestines. Epithelial tumors of the small intestine, cecum, and colon were analyzed for insertion sites.	Information for the CIS Start and CIS End fields was taken from Supplemental Table 4. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. Predicted effect information was submitted directly by the author.
Starr 2011-01	Starr et al. identified mutations that cooperate with mutant Apc to cause colorectal cancer by conducting a Sleeping Beauty mutagenesis screen on a C57BL6 ApcMin background using the villin promoter to activate transposition in intestines. Resulting epithelial tumors of the small intestine, cecum, and colon were analyzed for common insertion sites.	Information for the CIS Start, CIS End, and Predicted Effect fields was taken from Table 2. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. The other gene data was submitted directly by the author.
Suarez-Cabrera 2017-01	Suarez-Cabrera, et al., performed an SB screen in mice to identify drivers of breast cancer on a heterozygous conditional p53 background (Trp53F2-10). The SB system consisted of T2/Onc2 and a K5-SB11 transgene. The heterozygous p53 allele was generated using a K5-Cre transgene. Transposon insertions were analyzed in 33 mammary gland tumors from 26 mice. The majority of tumors were classified as acinar carcionma or adenocarcinoma with fewer papillary carcinoma or cystadenocarcinoma. The majority of tumors were ER positive. CISs were identified by gCIS and had to present in at least three separate tumors.	Data was taken from Table 1. Gene symbols and gneomic coordinates for GRCm38/mm10 were obtained using MGI Batch Query. Relative Rank was based on number of tumors in each CIS.
Takeda 2015-01	Takeda et al., performed three Sleeping Beauty screens to identify genes and gene networks that play a causal role in GI tract tumorigenesis. Tumors were a mixture of adenomas and adenocarcinomas. The three screens used three different pre-disposing mutations (KrasLsL-G12D, SMAD4+/-, Trp53floxedR217H). The model used the T2/Onc2 transposon, a conditional transposase (Rosa26-LsL-SB11) and an intestinal specific inducible Cre (Vil-CreERT2). Animals were given 4-OHT ~3 months after birth. The authors also performed the same screen using wild-type mice and ApcMin mice but did not publish the CIS lists for these screens. This screen (2015-01) contains the CIS genes found in the KrasLsL-G12D screen.	Data was taken from Table 2 and from an additional spreadsheet kindly provided by the author, Dr. Takeda, containing the GKC output. Genome coordinates were converted from NCBI37/mm9 to GRC38/mm10 using UCSC LiftOver. Gene symbols were converted to official RefSeq symbols. Relative rank was determined using approximate_p_value_adjusted_across_genome.
Takeda 2015-02	Takeda et al., performed three Sleeping Beauty screens to identify genes and gene networks that play a causal role in GI tract tumorigenesis. Tumors were a mixture of adenomas and adenocarcinomas. The three screens used three different pre-disposing mutations (KrasLsL-G12D, SMAD4+/-, Trp53floxedR217H). The model used the T2/Onc2 transposon, a conditional transposase (Rosa26-LsL-SB11) and an intestinal specific inducible Cre (Vil-CreERT2). Animals were given 4-OHT ~3 months after birth. The authors also performed the same screen using wild-type mice and ApcMin mice but did not publish the CIS lists for these screens. This screen (2015-02) contains the CIS genes found in the SMAD4+/- screen.	Data was taken from Table 2 and from an additional spreadsheet kindly provided by the author, Dr. Takeda, containing the GKC output. Genome coordinates were converted from NCBI37/mm9 to GRC38/mm10 using UCSC LiftOver. Gene symbols were converted to official RefSeq symbols. Relative rank was determined using approximate_p_value_adjusted_across_genome.
Takeda 2015-03	Takeda et al., performed three Sleeping Beauty screens to identify genes and gene networks that play a causal role in GI tract tumorigenesis. Tumors were a mixture of adenomas and adenocarcinomas. The three screens used three different pre-disposing mutations (KrasLsL-G12D, SMAD4+/-, Trp53floxedR217H). The model used the T2/Onc2 transposon, a conditional transposase (Rosa26-LsL-SB11) and an intestinal specific inducible Cre (Vil-CreERT2). Animals were given 4-OHT ~3 months after birth. The authors also performed the same screen using wild-type mice and ApcMin mice but did not publish the CIS lists for these screens. This screen (2015-03) contains the CIS genes found in the p53floxedR217H screen.	Data was taken from Table 2 and from an additional spreadsheet kindly provided by the author, Dr. Takeda, containing the GKC output. Genome coordinates were converted from NCBI37/mm9 to GRC38/mm10 using UCSC LiftOver. Gene symbols were converted to official RefSeq symbols. Relative rank was determined using approximate_p_value_adjusted_across_genome.
Takeda 2016-01	Takeda, et al., performed an SB screen to identify genes that cooperate with loss of SMAD4 to cause gastric cancer. They combined a germline heterozygous SMAD4 knockout with a Rosa26-LsL-SB11 knockin and a T2/Onc3 concatamer (11x on chromosome 9). SB11 was activated using a ACTB-Cre transgene. Mice developed tumors throughout the body, but the current study only analyzed 66 gastric tumors from 35 mice.	Data was extracted from supplemental tables 2, 3 and 6. Genome coordinates for GKC CISs (Supplemental Table 3) were converted to mm10 using UCSC LiftOver. Genome coordinates identified only by GCIS were based on the genomic coordinates of the gene, extracted from MGI batch query. Relative rank was based on % of tumors mutated, with the highest percentage used if identified in both GKC and GCIS.
Tang 2013-01	Tang et al. sought to define the genetic lesions driving leukemia by directing Sleeping Beauty mutagenesis toward the blood-forming system in mice. Mice homozygous for the T2/Onc2 transposon concatemer and cre-inducible RosaSBase were crossed with mice expressing cre from the oncogenic, hematopoietic-specific vav1 promoter. Two screens were conducted. In the first, a transgene expressing JAK2V617F was also expressed in hematopoietic cells. In the second, SB mutagenesis was repeated in leukemic mice harboring hematopoietic cells expressing translocated in liposarcoma (TLS)-ERG. Linker-mediated PCR on DNA extracted from spleen cells revealed integration sites. Statistically significant common insertion sites were identified by GKC analysis.	Data for screen 01 were gathered from Table 1. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. No predicted effect information was included in the data.
Tang 2013-02	Tang et al. sought to define the genetic lesions driving leukemia by directing Sleeping Beauty mutagenesis toward the blood-forming system in mice. Mice homozygous for the T2/Onc2 transposon concatemer and cre-inducible RosaSBase were crossed with mice expressing cre from the oncogenic, hematopoietic-specific vav1 promoter. Two screens were conducted. In the first, a transgene expressing JAK2V617F was also expressed in hematopoietic cells. In the second, SB mutagenesis was repeated in leukemic mice harboring hematopoietic cells expressing translocated in liposarcoma (TLS)-ERG. Linker-mediated PCR on DNA extracted from spleen cells revealed integration sites. Statistically significant common insertion sites were identified by GKC analysis.	Data for screen 02 were gathered from Table 02. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. No predicted effect information was included in the data.
Van der Weyden 2011-01	Van der Weyden et al. used Sleeping Beauty mutagenesis screens to identify candidate cancer genes driving B-cell acute lymphoblastic leukemia (BCP-ALL) in conjunction with the ETV6-RUNX1 fusion gene, a hallmark of BCP-ALL. The endogenous Etv6 allele was used to drive expression of a floxed cassette containing the ETV6-RUNX1 fusion gene and the HSB5 transposase, separated by an IRES. Mice expressing this construct were crossed with mice harboring the T2/Onc2 transposon array to drive transposition. Resulting leukemic tissue was used for DNA extraction and subsequent splinkerette PCR, followed by sequencing. GKC analysis pinpointed statistically significant common insertion sites.	All data were gathered from Supplementary Table S3. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. The listed peak location for each CIS was used as both the Start and End coordinate. Data only included insertions for each CIS, and not unique insertions. Two CISs within the Mecom gene were listed separately, so the CIS with the higher insertion count was included in the CCGD. No predicted effect information was included in the data.
Van der Weyden 2012-01	Van der Weyden et al. studied the effect of Cadm1 loss in promoting cancer and the commonly mutated genes exacerbating this effect using Sleeping Beauty mutagenesis. Transposition of the T2/Onc concatemer was driven by the SB11 transposase at the Rosa26 locus in both Cadm1-/- and Cadm1+/+ mice. Leukemias and lymphomas were the most commonly formed tumor types and were therefore used for the determination of common insertion sites. Lymphomatous tissue DNA was extracted and splinkerette PCR, followed by sequencing, identified insertion sites. GKC analysis was used to narrow insertion sites down to a list of the statistically most common.	All data were gathered from Figure 4. Limited CIS data provided. CIS regions identified with a genome-wide adjusted P-value of >0.05 were excluded. No genome build information was provided, so for the genes with a peak location listed, coordinates were converted from NCBI37 build to GRCm38/mm10, using the peak as both the Start and End value. For CIS regions lacking a peak location, the Start and End coordinates correspond to the limits of the gene. Screen 01 corresponds to CIS regions from Cadm1+/+ mice. No predicted effect information was included in the data.
Van der Weyden 2012-02	Van der Weyden et al. studied the effects of Cadm1 loss in promoting cancer and the commonly mutated genes exacerbating these effects using Sleeping Beauty mutagenesis. Transposition of the T2/Onc concatemer was driven by the SB11 transposase at the Rosa26 locus in both Cadm1-/- and Cadm1+/+ mice. Leukemias and lymphomas were the most commonly formed tumor types and were therefore used for the determination of common insertion sites. Lymphomatous tissue DNA was extracted and splinkerette PCR, followed by sequencing, identified insertion sites. GKC analysis was used to narrow insertion sites down to a list of the statistically most common.	All data were gathered from Figure 4. Limited CIS data was provided. CIS regions identified with a genome-wide adjusted P-value of >0.05 were excluded. No genome build information was provided, so for the genes with a peak location listed, coordinates were converted from NCBI37 build to GRCm38/mm10, using the peak as both the Start and End value. For CIS regions lacking a peak location, the Start and End coordinates correspond to the limits of the gene. Screen 02 corresponds to CIS regions from Cadm1-/- mice. No predicted effect information was included in the data.
Van der Weyden 2013-01	Van der Weyden et al. aimed to determine which somatic gene mutations can drive leukaemogenesis in the context of Trp53 heterozygosity using Sleeping Beauty mutagenesis. The T2/Onc transposon concatemer was used for mutagenesis, activated by Rosa26SB11, in mice heterozygous, homozygous, or wildtype for the mutant Trp53Tyr allele. Genomic DNA was extracted from leukaemic/lymphomic tissue in moribund mice, which most commonly formed in the spleen, thymus, or lymph node. Linker-mediated PCR and sequencing identified points of transposon insertion. GKC analysis revealed the statistically common insertion sites.	All data were gathered from Supplementary Table S1. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. Three screens were conducted. Screen 01 corresponds to Trp53+/+ mice. No predicted effect information was included in the data.
Van der Weyden 2013-02	Van der Weyden et al. aimed to determine which somatic gene mutations can drive leukaemogenesis in the context of Trp53 heterozygosity. The T2/Onc transposon concatemer was used for mutagenesis, activated by Rosa26SB11, in mice heterozygous, homozygous, or wildtype for the mutant Trp53Tyr allele. Genomic DNA was extracted from leukaemic/lymphomic tissue in moribund mice, which most commonly formed in the spleen, thymus, or lymph node. Linker-mediated PCR and sequencing identified points of transposon insertion. GKC analysis revealed the statistically common insertion sites.	All data were gathered from Supplementary Table S1. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. Three screens were conducted. Screen 02 corresponds to Trp53+/- mice. No predicted effect information was included in the data.
Van der Weyden 2013-03	Van der Weyden et al. aimed to determine which somatic gene mutations can drive leukaemogenesis in the context of Trp53 heterozygosity. The T2/Onc transposon concatemer was used for mutagenesis, activated by Rosa26SB11, in mice heterozygous, homozygous, or wildtype for the mutant Trp53Tyr allele. Genomic DNA was extracted from leukaemic/lymphomic tissue in moribund mice, which most commonly formed in the spleen, thymus, or lymph node. Linker-mediated PCR and sequencing identified points of transposon insertion. GKC analysis revealed the statistically common insertion sites.	All data were gathered from Supplementary Table S1. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. Three screens were conducted. Screen 03 corresponds to Trp53-/- mice. No predicted effect information was included in the data.
van der Weyden 2015-01	Van der Weyden, et al., performed an Sleeping Beauty screen in mice expressing an ETV6-RUNX1 fusion gene combined with heterozygous loss of Pax5. The transpoase (HSB5) was expressed from an IRES connected to ETV6-RUNX1. The transposon was T2/Onc2. Mice developed B-ALL, AML, T-ALL, and unclassified leukemia. 20 B-ALL tumors were sequenced using 454 technology and CISs were defined using Gaussian Kernel Convolution (GKC).	Data was taken fromFigure 2A. Relative Rank was determined using the Unique insertions in CIS column. CIS coordinates are the gene coordinates.
Vyazunova 2014-01	Vyazunova et al., performed an SB screen to identify genes that contribute to glioma formation. Gliomas arising in several different transgenic backgrounds were analyzed together for this study. Mice contained concatamers of transposons (T2/Onc, T2/Onc2, or T2/Oncatg), a constitutive transposase allele (Rosa26-SB11 or GFAP-SB11), and were either on a wild-type background or a predisposing background (p19ARF-/-, p19ARF+/-, Blm-/-, Blm+/-, or CSF1+/-). A total of 14 gliomas were analyzed, and CISs were calculated using a custom approach as well as a gene-centric approach.	Data was taken from Table 3 and Table S3. Genomic locations (based on CIS calls) were converted from NCBI37/mm9 to GRC38/mm10 using UCSC LiftOver. Relative rank was determined using number of tumors. Due to low diversity of tumor numbers, there were no ranks of D in this set.
Wong 2014-01	Wong, et al., performed a forward genetic screen in mice using Sleeping Beauty (CAGGS-SB11, T2/Onc, pIpC-induced Mx1-Cre) that produced T-cell lymphoblastic lymphoma (T-ALL). They analyzed insertion sites in 44 tumors and identified 90 genes.	Data was taken from Supplemental Table 4. Genome coordinates were converted from NCBI37/mm9 to GRCm38/mm10 using UCSC LiftOver. Relative rank is based on total number of insertions.
Wu 2012-01	Wu et al. utilized Sleeping Beauty transposon mutagenesis to study genes important for medulloblastoma formation and metastasis. Because approximately 30% of mice heterozygous for the PATCHED gene develop cerebellar medulloblastomas, the authors employed SB transposition in Ptc+/- mice. To accomplish this, they expressed the SB11 transposase in cerebral progenitor cells using the Math1 enhancer/promoter, driving site-specific integration of the T2/Onc concatemer. Mutations in the Tp53 gene are also associated with increased risk of medulloblastoma formation, so a second screen was conducted in a Tp53mut background using identical transposase/transposon constructs. Linker-mediated PCR on genomic DNA extracted from medulloblastomas amplified transposon-genomic junctions and sequencing identified integration sites.	All data were gathered from Supplementary Table S2. Two screens were performed. Screen 01 corresponds to Ptch+/- mice. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. Gene-centric analysis identified CIS regions. The Start and End coordinates correspond to the start and end locations of the gene containing the insertions. No predicted effect information was included in the data.
Wu 2012-02	Wu et al. utilized Sleeping Beauty transposon mutagenesis to study genes important for medulloblastoma formation and metastasis. Because approximately 30% of mice heterozygous for the PATCHED gene develop cerebellar medulloblastomas, the authors employed SB transposition in Ptc+/- mice. To accomplish this, they expressed the SB11 transposase in cerebral progenitor cells using the Math1 enhancer/promoter, driving site-specific integration of the T2/Onc concatemer. Mutations in the TP53 gene are also associated with increased risk of medulloblastoma formation, so a second screen was conducted in a Tp53mut background using identical transposase/transposon constructs. Linker-mediated PCR on genomic DNA extracted from medulloblastomas amplified transposon-genomic junctions and sequencing identified integration sites.	All data were gathered from Supplementary Table S2. Two screens were performed. Screen 02 corresponds to P53mut mice. Start and End coordinates were originally reported using NCBI37 build. They were converted to GRCm38/mm10 using UCSC LiftOver. Gene-centric analysis identified CIS regions. The Start and End coordinates correspond to the start and end locations of the gene containing the insertions. No predicted effect information was included in the data.
Wu 2016-01	Wu, et al., performed an SB screen to identify genes driving development of neurofibromas. Their model used a desert hedgehog Cre (Dhh-Cre) to initiate transposition and loss of NF1 (NF1fl/fl) in Schwann cells. The SB system consisted of Rosa26-LsL-SB11 and T2/Onc transposons. Insertion sites from 49 neurofibromas were analyzed using Monte Carlo criteria for statistical significance.	Data was extracted from Figure 1A. Official symbols and genomic coordinates from GRCm38/mm10 were obtained from MGI batch query. Relative rank was based on number of Neurofibromas with an insertion. CIS coordinates were not published.
Xu 2017-01	Xu, et al., performed a screen for liver cancer drivers by delivering a library of CRISPR/Cas9 sgRNAs using a Piggybac transposon. They used the GeCKOv2 library targeting all mouse protein coding genes and microRNAs. They targeted the liver by using high pressure tail vein injections. Mice were sensitized to liver cancer by co-delivering PB transposons containing sgRNA to Cdkn2a and a mutated RAS (hNRASG12V). All mice were euthanized at 45 days and the majority of tumors were Intrahepatic cholangiocarcinomas. 18 tumors from 8 mice were analyzed by deep sequencing for presence of sgRNAs. If an sgRNA sequence was detected at a high read count, the targeted gene was considered a candidate cancer tumor suppressor gene.	Data was extracted from Dataset 1. Official gene symbols and genomic coordinates for GRCm38/mm10 were obtained using MGI batch query. No relative rank was assigned due to the vast majority of sgRNAs being identified in only a single tumor.
Zanesi 2013-01	Zanesi et al. investigated commonly mutated genes cooperating with overexpressed Tcl1 in B-cell chronic lymphocytic leukemia (CLL) using Sleeping Beauty mutagenesis. A CD19-Cre transgene activated transposition of the SB11 transposase at the Rosa26 locus specifically within B-cells. The SB6070 and SB6113 strains harboring the T2/Onc2 transposons in different locations were both used. Linker-mediated PCR on DNA extracted from tumorigenic mice, followed by gene-centric analysis, defined common insertion sites.	All data were gathered from Table 1. Because gCIS analysis was used, the insertion coordinates listed in the CCGD correspond to the start and end points of the genes containing the CIS. All coordinates correspond to the GRCm38/mm10 annotation. Predicted effect information was reported in Table 1 by RT-PCR of CIS genes. CISs resulting in higher expression of affected genes in > 50% of tumor samples were designated as Gain. All predicted effect data was listed in Table 1.