Equipo completo

2.10.1 Library preparation

WES was carried out using commercial kits from Agilent Technologies: SureSelect XT Library Preparation kit ILM, SureSelect Target Enrichment and Herculase II Fusion DNA Polymerase, SureSelect XT Human All Exon V4 Capture or SureSelect XT Human All Exon V5 Capture Libraries. The standard Illumina protocol “SureSelect XT Target Enrichment System for Illumina Paired-End Sequencing Library” was followed (available from http://www.agilent.com/cs/library/usermanuals/ Public/G7530- 90000_SureSelect_IlluminaXTMultiplexed_1.8.pdf). Briefly, each DNA sample was quantified using the Qubit® dsDNA (Broad-Range) Assay (Invitrogen) (Section 2.3). 3μg of gDNA was diluted in 1 x TE buffer to a final volume of 250μl and added to a clearly labelled T6-30 glass tube (Covaris, USA). The DNA was then sheared in a water bath at 20°C to fragment sizes of 150 to 200 bp using the Covaris system (Covaris S220 Sonicator and SonoLite software). This sheared DNA sample was then purified using Agencourt AMPure XP magnetic beads (Beckman Coulter Genomics, South Plainfield, USA) and analysed using a DNA 1000 BioanalyzerTM _{(Agilent Technologies) to assess the} distribution of DNA fragment sizes between 150-200 bp.

The creation of blunt-ended fragments and 5'-phosphorylation of the ends were the next steps. 48µl of the purified sample was mixed with 35.2µl of nuclease-free water, 10µl of 10 x end repair buffer, 1.6µl of dNTP mix, 1µl of T4 DNA polymerase, 2.2µl of T4 Polynucleotide Kinase, and 2 µl Klenow DNA Polymerase. This mixture was incubated in a thermal cycler for 30 minutes at 20°C, then the DNA fragments were again purified using Agencourt AMPure XP magnetic beads. After the fragment ends were repaired, adenosine overhangs were added to their 3’ ends (3'-dA overhangs) by incubation of 30µl of the sample with 3 µl of Exo(-) Klenow, 5µl of 10 × Klenow polymerase buffer, 11µl of nuclease-free water and 1µl dATP for 30 minutes at 37°C.

The sample was then purified a third time using the magnetic beads. Paired-end adaptors were ligated on by adding 1.5µl T4 DNA Ligase, 10µl of 5 x T4 DNA ligase

55

buffer, 10µl of diluted SureSelect adaptor oligo mix and 15.5µl of nuclease-free water to 13μl of the DNA sample, followed by incubating for 15 minutes at RT. Next, the sample was purified using the magnetic beads again and the adapter-ligated library was amplified in a 50μl PCR containing, 15μl of DNA sample, 1.25µl of SureSelect ILM indexing pre- capture PCR reverse primer, 1.25µl SureSelect primer, 10μl of 5 x Herculase II reaction buffer, 1 μl of Herculase II fusion DNA polymerase, 0.5μl of 100mM dNTP mix, and 21μl of nuclease-free water. The PCR mix was loaded into a thermal cycler with the following program: 98°C for 2 minutes (denaturation) and 6 cycles of 98°C for 30 seconds, 65°C for 30 seconds and 72°C for 1 minute, followed by a final extension step of 72°C for 10 minutes. The amplified library was purified using the magnetic beads and analysed using a DNA 1000 BioanalyzerTM assay. Only samples with an electropherogram reading showing a single peak around 250 to 275 bp were taken through to the hybridisation steps.

Each amplified library was quantified using the PicoGreen® double stranded DNA (dsDNA) quantitation assay (Invitrogen) according to the manufacturer’s instructions. The library was then hybridized and the exome was captured individually prior to addition of the indexing tag. For each hybridization reaction, 40μl of the hybridization buffer and 5.6μl of SureSelect block mix were prepared according to the manufacturer’s protocol. 750ng of the library DNA with a maximum volume of 3.4μl was added to 5.6μl of prepared SureSelect block mix then incubated for 5 minutes at 95°C. In a PCR plate maintained at 65°C, the prepared library was mixed with 13μl of the prepared hybridization buffer and 7μl of SureSelect capture library of biotinylated RNA oligonucleotide probes (5μl of SureSelectXT Human All Exon V4/V5 and 2μl of 25% RNase block), followed by incubation for 24 hours at 65°C in a thermal cycler with a heated lid at 105°C. The captured library was fished out using streptavidin-coated magnetic beads (Dynabeads MyOne Streptavidin T1, Invitrogen) on a magnetic separator according to the protocol guidelines. Subsequently, index tags were added to the captured library by post-hybridization amplification. 14μl of each DNA sample and 1μl of the appropriate index PCR primer were mixed with 35μl of Herculase II master mix (10μl of 5 x Herculase II reaction buffer, 1μl of Herculase II fusion DNA polymerase, 0.5μl of 100mM dNTP mix, 22.5μl of nuclease-free water and 1μl of SureSelect ILM Indexing post capture forward PCR primer), followed by loading into a thermal cycler for 2 minutes at 98°C, then 12 cycles of 30 seconds at 98°C, 30 seconds at 57°C and 1 minute at 72°C.

56

Final extension was at 72°C for 10 minutes. The samples were then purified using Agencourt AMPure XP beads and analysed using the 2100 BioanalyzerTM high sensitivity DNA assay (Agilent Technologies) which was expected to achieve a normal distribution around a peak ranging from approximately 300 to 400 bp. Six samples were pooled together in a final volume of 50μl, with each sample having a final concentration of 10nM. Finally, the cluster amplification was performed at the NGS facility (University of Leeds), followed by NGS using paired-end 100 bp reads on an Illumina 2500 HiSeq sequencer (Illumina Inc. UK).

2.10.2 Analysis of WES data.

The computational analysis of the WES data was performed using Unix console commands (Appendix 1) and a wide range of online servers. The quality of the raw data coming from the Illumina high throughput sequencing was determined by using FASTQC tools run on the Galaxy platform (Blankenberg et al., 2010). Quality scores across all bases, GC content per sequence, sequence length distribution and duplication levels were evaluated before any further analysis. After sequence quality monitoring, the sequencing data was aligned against the reference genome (hg19/GRCh37) using either the Bowtie2 program (Langmead and Salzberg, 2012) or NovoAlign software (http://www.novocraft .com/ products/novoalign/). NovoAlign was preferred in exome depth and fishing CNV analysis or filtering the sample against in-house samples that had been analysed using the same aligner. Otherwise the Bowtie2 aligner was more widely used. The aligned files were sorted, indexed and processed in SAM/BAM format using the SAMtools suite of programs (Li et al., 2009) (http://samtools. sourceforge.net/). PCR duplicates were removed by Picard tools (http://broadinstitute. github.io/ picard/). The mean depth of reads per base was observed and the variants were then realigned locally and recalibrated using the Genome Analysis Toolkit (GATK) (https://www. broadinstitute. org/gatk, version 3.3-0). Indel and single nucleotide variants were called in the variant call format (VCF) format using the Unified Genotyper function of GATK (DePristo et al., 2011). The Integrative Genomics Viewer (IGV) (https://www. broadinstitute.org/igv/) (Robinson et al., 2011) was used for visualization and interactive exploration of the aligned data files. The variant list of each individual was annotated using ANNOVAR software (http://annovar. openbioinformatics.org/en/latest/) and filtered using the following criteria: selecting only DNA variants in coding regions, splice donor and

57

acceptor sites (±2 bp), removing synonymous changes, selecting variants with a greater than depth of coverage of 10 reads, filtering out exome variants that have a minor allele frequency (MAF) of greater than 1% in either dbSNP 138 (http://www.ncbi. nlm.nih.gov/SNP/), the 1000 Genomes (Abecasis et al., 2012), the exome variant server (EVS) or the Exome Aggregation Consortium (ExAC) databases (Section 2.14.4). For VCF files of patients of Pakistani origin, an additional filtering step was used to filter out all variants with MAF greater than 5% in a cohort of 3222 exomes of British Pakistani adults (Narasimhan et al., 2016) (http://www.genesandhealth.org/research/scientific- data-downloads), this step was used through a perl script (Appendix 1) developed by Dr David Parry, University of Leeds. The final variant lists were initially compared to the known retinal dystrophy genes in the RetNet database (URL: https://sph.uth. edu/retnet/). The pathogenicity of variants was also assessed using a number of software either integrated in the ANNOVAR or used separately (Section 2.14.2).