T.2 Sequencing double stranded DNA on AB

A polyacrylamide gel was prepared. First the plates (35 x 24 cm) were washed in hot water only, rinsed with distilled deionized water, dried with one wipe of a white kimwipe and assembled horizontally with 0.4 mm spacers. In a clear plastic container the 50 ml of the 6% acrylamide, 8.3M urea, 1 x TBE gel mix was made as follows:

40 ml Sequagel 6 (National Diagnostics)

10 ml Sequagel Complete (which contains TEMED)

400 pi 10% ammonium persulphate (APS) was added and the solution sucked into a syringe for pouring. It was poured and a comb placed. The gel was allowed 40 minutes to set.

The ABI machine (ABI 373A DNA Sequencer, ABI, Warrington, Cheshire, UK) was set for filter set A and the PMT (photomultiplier tube) voltage adjusted to be within 800-

When the machine was set the comb was removed and the plates were positioned in the machine and scanned to ensure that they were clean. If this was satisfactory the plate- heater and acrylic bar were added, the buffer tank fixed, top and bottom buffer tanks filled with 1,000 ml 1 x TBE and the electrodes connected. A pre-run of 15-60 mins was carried out. Meanwhile the necessary settings for the run (10 hours, 2500 V, 40mA, 25W) and soft ware for data collection and subsequent analysis was set up.

Immediately prior to loading the gel a 24-well sharks-tooth comb was positioned and the samples were heated to 94°C for 3 mins to denature them. They were loaded into

alternate wells, run in for 5 mins and then the remaining wells loaded.

After the 10 hour run the sequence data was stored on an optical disc drive and analysed. Analysis of the data was performed using an Apple Macintosh computer with software linked to the sequencer (ABI Prism). The output consisted of two files, a text only file, containing the read sequence, and an analysis file containing the raw data, analysed data and detailed sequence information. The quality of the sequence could be judged by the electropherogram or sequence scan presented within this file.

2.2.8 Computer-aided analysis 2.2.8.1 Genotyping

The developed X-ray films with the radiolabelled microsatellite data were analysed on a light box. Alleles for the marker were identified as a system and assignments made for each individual. The alleles were then read across the pedigree and the number of cross­ overs counted. In this way it was possible to check that the alleles fitted the pedigree. The allelic frequencies were taken from HGMP if the family was small and did not contain more than 10 spouses, but otherwise the allele frequencies were estimated from the spouses.

2.2.8.2 Linksys

Allele information was stored on a SUN Microsystems mainframe computer under a UNIX operating system. The linkage analysis was run on this computer via a local area network (HGMP-RC, Harrow, Middx). Some markers were analysed using Cyrillic software (using MLINK, see below). Each pedigree and its disease allocation was entered and copies were made of these family files, because information on only 9 marker loci could be stored in each family file. Genotype information for each marker was then entered on each individual sampled. Locuslib and phenolib files containing the information on the phenotype and microsatellite marker characteristics were made

(LINKSYS) (Attwood and Bryant, 1988). FED and DAT files were created and exported to the LINKAGE sub directory.

2.2.5.3 Two-point linkage

Pedigree data was formatted by MAKEPED subprogram. Two-point Lod scores were calculated by MLINK (Lathrop, Lalouel, Julier et al., 1984). The optic atrophy pedigrees were analysed as an autosomal dominant trait, with a penetrance of 0.98, equal allelic- frequencies, juvenile onset, no phenocopy rate and a gene frequency of 0.0001. ILINK was used to generate maximum Lod scores.

2.2.5.4 Genetic homogeneity analysis (HOMOG)

The A-test {admixture test) for heterogeneity assumes that there are two types of families.

One type with linkage (q = qi<0.5) and one without (q = 0.5). a is the proportion of linked families, i.e., q = q i in proportion a of the families, and q = 0.5 in proportion (1 - a) of the families, q is the recombination fraction in the linked families. The likelihood of any given family showing heterogeneity is:

L(a, q) = a L(q) + (1 - a)L(q = 0.5)

The calculation is carried out using the HOMOG program of the LINKAGE package (Ott, 1991) A test for genetic heterogeneity/ homogeneity was applied to five of the larger families, pedigrees A l, G l, K l, LI and W1 using HOMOG version 3.3. Lod scores at recombination fractions (theta) from 0.00 to 0.45 were computed in steps of 0.05 for the disease locus flanking markers. One HOMOG analysis was carried out for the five families at each of 5 flanking markers.

For the input file N = the number of values at which the Lod scores are given (i.e., 10), STEPSIZE = the step size for increments of a (i.e., 0.05), OUT = 0, ALOW = 0, LL = 80, 91, 92, 9N are the values of theta for which the lod scores are provided (i.e., 0.00 to 0.45) and NFAM = the number of families for which the Lod scores are provided (i.e., 5). A Lod score of -oo is represented as -80.

The three hypotheses being tested were: H2: Linkage, heterogeneity

HI: Linkage, homogeneity HO: No linkage

The three likelihood ratio tests possible with these three hypotheses were: H2 vs. HI = the test of heterogeneity given linkage

HI vs. HO = the standard test of linkage assuming homogeneity H2 vs. HO = the joint test of linkage and heterogeneity.