We hypothesize that normal germline variation in gene expression accounts, in part, for variation in the clinical severity and phenotype of the monogenic disorder neurofibromatosis type 1 (NF1). The phenotype of NF1 is constituted by a variety of quantifiable features; we term these features "subphenotypes". Our main focus is on sub-phenotypes with published evidence of variation in expression from an inherited component. These include our primary sub-phenotype of spinal neurofibroma burden (quantified by MRI of spinal neurofibromas) and the secondary sub-phenotypes of dermal neurofibroma burden, head circumference, number of Lisch nodules, scoliosis, history of plexiform neurofibromas, and height. Other sub-phenotypes, as collected by a routine history and physical, will also be evaluated.

According to our hypothesis, the severity of a sub-phenotype will correlate with heritable differences in the germline expression of certain genes. As an example, consider a spectrum of individuals affected with the primary sub-phenotype of spinal neurofibroma burden. For most genes, there will be no relationship of expression level to the number of spinal neurofibromas. However, some genes will have a direct (or inverse, or other) correlation of their level of expression and the number of neurofibromas. Such genes would be considered as candidate modifier genes. The secondary sub-phenotypes will also be analyzed in a similar

way. To limit false positives, candidate genes will then be tested for association and linkage (using the transmission/disequilibrium test, TDT, and affected sibling pairs, ASPs) with the appropriate sub-phenotype.

Recruitment will be focused on identifying individuals with a range of severity of the primary sub-phenotype, spinal neurofibroma burden. We will primarily recruit post-pubertal individuals who meet the 1988 NIH criteria for neurofibromatosis type 1 (NIH Consensus Development Conference 1988). We will exclude those with recognized segmental or mosaic NF1. The rate of cutaneous neurofibroma growth and number is known to vary widely; these tumors typically appear in adolescence. For this reason we will ascertain patients after puberty. We expect most individuals to be 18 years or older, but will also accept post-pubertal pediatric patients (and use a hand film to demonstrate bone age). Parents (whether affected or not) are critical when using family-based test of association (like the TDT) and test of linkage and will also be recruited.

To detect linkage, a minimum of 200 ASPs with spinal neurofibromas is needed. Tests of association with the TDT require trios (mother, father, proband) that can be derived from ASP families. However, we will also recruit trios independent of ASP families. Thus, we set a recruitment goal of 1500 individuals (500 affected individuals, plus 1000 parent or additional sibs in about 400 families) to accommodate this requirement.