Whole exome sequencing of KrasG12D-driven mouse pancreatic cancer cell cultures derived from primary pancreatic tumors (n=38) and liver or lung metastases (n=5).

We generated primary pancreatic ductal adenocarcinoma (PDAC) cell cultures (38 primary pancreatic tumors and 5 corresponding liver or lung metastases) from mice expressing mutant KrasG12D conditionally in the pancreas (PK mice). Coding exons were enriched via whole-exome pull down using Agilent SureSelect Mouse Exon Kit according to manufacturer's instructions and sequenced on the Illumina HiSeq2000 system. Prior to mapping, raw sequencing reads were trimmed using Trimmomatic v0.33. Leading and trailing bases with phred scores below 25 and reads shorter 50 nucleotides were removed. In addition the average base quality within a sliding window of 10 nucleotides should be above 25 to keep the read for further downstream analysis. Reads were aligned to the GRCm38.p3 reference genome using BWA-MEM 0.7.12 with default settings. PCR duplicates were marked with Picard tools v1.130 and realignment around indels was performed with GATK toolkit v 3.4.46. Mutect v 1.1.7 was used for calling somatic mutations with default settings. Potential somatic events were filtered for single nucleotide polymorphisms (SNPs) by excluding SNVs which were listed in in release 1505 of the Mouse Genome Project SNP database. Somatic point mutations were included in the final list, if the read coverage for each position was =10 in both control and tumor, variant frequency was =10% and read count supporting the variant nucleotide is =3 in the tumor sample and =0 in the control. Further, SNVs marked as strand or PCR bias artifacts by "DKFZBiasFilter" (https://github.com/eilslabs/DKFZBiasFilter, using default settings) or with a FOXOG-Score of 1 were excluded. Annotation of somatic events was conducted with SNPeff v4.1. SNVs causing variation in splice sites or upstream/downstream of genes were excluded from further analysis. Indels were detected with Pindel. For each potential indel the read coverage was re-calculated using bedtools v2.17.0. Criteria for further downstream processing were: Variant frequency =10 in tumor and =0 in control; and total coverage at the altered position in both control and tumor =20.For LOH analysis variant positions in control and tumor were computed with samtools mpileup v1.3.1. Only positions in regions with mapping quality of 60 and an average phredscore of 20 were considered for further analysis. Furthermore, positions harboring strand bias and variant allele frequencies less than 20% and above 85% in the control were excluded as they are likely homozygous in the germline. The minimal cutoff coverage for a given polymorphic position in the control was set to eight reads. Segmental duplications (UCSC Genome Browser) and regions with mouse line specific variation (Mouse Genomes Project, REL-1505) were excluded. For this set of SNPs the difference of frequencies between tumor and control samples were calculated.