Additional file 2: of Analysis of error profiles in deep next-generation sequencing data MaXiaotu ShaoYing TianLiqing FlaschDiane MulderHeather EdmonsonMichael LiuYu ChenXiang NewmanScott NakitandweJoy LiYongjin LiBenshang ShenShuhong WangZhaoming ShurtleffSheila RobisonLeslie LevyShawn EastonJohn ZhangJinghui 2019 Supplementary Figures S1-S13. Figure S1. Comparison of mutant allele fraction (MAF) in diluted samples (y-axis) and undiluted cancer cell line (x-axis). Figure S2. Copy-number status of cell line COLO829 and ploidy of the 19 selected substitutions in this work. Figure S3. Quality metrics of sequenced datasets. Figure S4. Heatmap of error profiles across sequencing providers, sequencers, PCR enzymes, replicates, and dilutions. Figure S5. HiSeq error profile under CleanDeepSeq. Figure S6. Context dependency of C>T/G>A errors in HiSeq data under CleanDeepSeq. Figure S7. NovaSeq+Kapa error profile under CleanDeepSeq. Figure S8. Context dependency of C>T/G>A errors in NovaSeq+Kapa dataset under CleanDeepSeq. Figure S9. Context dependency of C>T/G>A errors in NovaSeq+Q5 dataset under CleanDeepSeq. Figure S10. False-postive introduced by “forced calling”. Figure S11. Error profiles in downsampling of NovaSeq + Q5 dataset. Figure S12. Comparison of standard pileup and CleanDeepSeq by using deepSNV on dilution experiments. Figure S13. Comparison of standard pileup and CleanDeepSeq by using MuTect on dilution experiments. (DOCX 5524 kb)