Instrument: 454 GS FLX
Strategy: Bisulfite-Seq
Source: GENOMIC
Selection: RANDOM
Layout: SINGLE
Construction protocol: DNA from individual CFU, CD34+, or RKO cells was harvested by Qiagen genomic DNA kit. DNA was next bisulfite modified and PCR amplified with forward oligonucleotides MLH1-AF (5' [Linker/spacer][barcode]actcaaaatcctctaccttataatatc 3'), MLH1-BF (5' [Linker/spacer][barcode]acaaaccaaacacaaaaccccat 3'), MLH1-CF (5' [Linker/spacer][barcode] tcacctcaacaaaaacacacaaac 3') when [linker/spacer]= (5' CGTATCGCCTCCCTCGCGCCATCAG 3') and [barcode] was sample specific. Negative oligonucleotides used were: MLH1-AR (5'[Linker/spacer][barcode]ttaaaagaagtaagatggaag 3'), MLH1-BR (5' [Linker/spacer][barcode]tttagttaataggagtagagatg 3'), MLH1-CR (5' [Linker/spacer][barcode]GTTAAATTTTTTAATTTTGTGGGTTGTTGGG 3') when [linker/spacer]= (5' CTATGCGCCTTGCCAGCCCGCTCAG 3') and [barcode] was sample specific. Amplified products were run on an 1.5% agarose gel and appropriate sized bands excized. DNA fragments were purified from gel fragments and submitted for 454 sequencing. In parallel 1/2 of each sample was processed for RNA extracts. Total RNA was generated for each sample and used as the template for cDNA synthesis by a single cycle of reverse transcriptase activity followed by RNA digestion. Expression of the human gene hMLH1 (in parallel with human beta actin as a control) was next assessed for each cDNA sample generated by QRT-PCR on an Applied biosystems RT-PCR 7500 fast machine. Threshold values were automatically generated according to the manufacturers protocol. Gene expression for hMLH1 (scored as 1) was considered observed if amplification products broke threshold for both hMLH1 and beta actin. However, loss of hMLH1 (scored as 0) gene expression was defined as the observation of detectable amplification products for beta actin but not hMLH1