Preclinical and cfDNA methylome profiling for detection and subtyping of Small Cell Lung Cancers

Methyl-Binding Domain protein sequencing (MBD-Seq) was applied to samples derived from patients with small cell lung cancer (SCLC), as well as non-cancerous controls. This included circulating tumour cell derived explants (CDXs) or patient derived Xenograft (PDX) models derived from 33 patients with SCLC, circulating cell-free DNA (cfDNA) from 78 patients with SCLC, cfDNA from 79 non-cancer controls and 13 non-cancerous lung tissue samples. The objects deposited here are R data files containing qseaSets from the R package qsea, which includes the read counts per sample per 300 base pair window across the genome, as well as information on copy number variation and metadata tables.<br> <br> Details of files: <strong>DX_All_min50_max1000_w300_q10.rds</strong> A qseaSet containing all 97 CDX/PDX samples (including replicates) and the 13 normal lung tissue samples. Note that min50_max1000_w300_q10 refers to including paired reads with between 50 and 1000 base pairs (bp), a window size of 300bp and a minimum MAPQ score of 10. <strong>DX_merged.rds</strong> A qseaSet with the biological replicates of each CDX merged together (formed from the above dataset). <strong>cfDNA_All_min50_max1000_w300_q10.rds</strong> A qseaSet containing 157 cfDNA samples used in the main body of the paper. <strong>NCCsForTrain_All_min50_max1000_w300_q10.rds</strong> A qseaSet containing the 38 NCC cfDNA samples used in the mixture sets for training the tumour/normal classifier. These samples are a subset of those in the cfDNA_All_min50_max1000_w300_q10.rds object. <strong>ValidationSet_All_min50_max1000_w300_q10.rds</strong> A qseaSet containing the 41 NCC cfDNA samples and 78 SCLC cfDNA samples used to validate the tumour/normal classifier. A subset of the cfDNA_All_min50_max1000_w300_q10.rds object. <strong>cfDNApostTreatment_All_min50_max1000_w300_q10.rds</strong> A qseaSet containing 7 cfDNA samples which were collected at a later, post-treatment, timepoint (mostly disease progression) from the same patients as in the main cfDNA object. Used only for Extended Figure 5, not any other part of the manuscript. <strong>varyDNAinput_All_min50_max1000_w300_q10.rds</strong> A qseaSet containing independent replicates of the cell line H1975, with different ng amounts of starting DNA (1-75ng). Used only for Figure 1B. <strong>TNmixSets_combined_regionsFiltered.rds</strong> A qseaSet containing the synthetic mixture sets generated by mixing either a CDX/PDX sample and a NCC sample or two NCC samples. Used to train the tumour/normal classifier. Restricted to only the windows used in the classifier for size regions, but the original mixtures are across the whole genome. <strong>CDXarrayWide.csv</strong> Pre-processed 450k Infinium Methylation array beta values for 8 CDX samples which were previously sequenced. Used for Supplementary Table 7 only. <strong>ArrayCDXs_percent100.rds</strong> A qseaSet containing the 8 CDX samples processed on 450k Infinium Methylation arrays (CDXarrayWide.csv) converted to estimated reads. <strong>KeyTFsincYAP.csv</strong> Variance stabilised transform (vst) values generated from RNASeq for the CDX/PDX samples, for the key genes involved in the subtype classifications. <strong>infinium-methylationepic-v-1-0-b5-manifest-file.csv</strong> A lookup file for the Infinium EPIC arrays, as downloaded from https://emea.support.illumina.com/downloads/infinium-methylationepic-v1-0-product-files.html. <strong>SCLC_transcript_expression_adjusted_for_batch_effects.csv</strong> SCLC_methylation_beta_values_of_individual_probes_after_QC_and_filtering_out_SNVs.csv Pre-processed transcript and methylation beta values from Infinium EPIC arrays for SCLC cell lines, as downloaded from sclccelllines.cancer.gov/sclc/downloads.xhtml (data timestamped as December 2019). <strong>CellLine_100percent.rds</strong> A qseaSet containing the SCLC cell lines converted to estimated MBD-Seq reads. <strong>DilutionSeries_CDX13_min50_max1000_w300_q10.rds</strong> <strong>DilutionSeries_CDX29_min50_max1000_w300_q10.rds</strong> <strong>DilutionSeries_CDX32_min50_max1000_w300_q10.rds</strong> Three qseaSets containing the results of an in silico dilution of a CDX (CDX13 = POU2F3, CDX29 = NEUROD1, CDX32 = ASCL1) with a single NCC, used to test limit of detection of the subtype classifier. Reads were mixed at the fastq level, prior to NextFlow pipeline being used. <strong>DilutionSeries_H446_Rep*_min50_max1000_w300_q10.rds</strong> Eleven qseaSets containing the results of an in silico dilution of a SCLC cell line H446 with a single NCC (not used to build the classifier), used to determine limit of detection of the tumour/normal classifier. Reads were mixed at the fastq level with different random seeds, prior to a NextFlow pipeline being used.