Determining a probabilistic diagnosis of autism by analysis of genomic copy number variations

A method for determining whether a patient is autistic, comprising:

• (a) obtaining a segmented genomic profile of DNA extracted from a cell of the patient, the segmented genomic profile comprising information about the copy number of a plurality of discrete segments within one or more genomic regions, the one or more genomic regions having been determined by:
• (i) obtaining genomes from a plurality of individuals that have autism;
• (ii) obtaining genomes from a plurality of individuals that do not have autism;
• (iii) identifying one or more genomic regions where copy number polymorphisms in the genomes of individuals that do not have autism are rare relative to copy number polymorphisms in individuals that have autism;
• (b) determining, by computer analysis, a genomic perturbation index value from the segmented genomic profile, wherein said perturbation index value incorporates:
• (i) the number of said discrete segments,
• (ii) the lengths of said discrete segments, or
• (iii) the distribution of the lengths of at least two adjacent segments;
• (c) comparing the perturbation index value of the patient to a perturbation index value determined from a segmented genomic profile of an individual known to have autism, the segmented genomic profile of the individual known to have autism comprising information about the copy number of a plurality of discrete segments within the same one or more genomic regions identified by the process in step (a);
• and
• (d) classifying the patient as autistic or non-autistic based on the comparison of step (c).