During the past decade numerous studies have been published describing chromosomal regions potentially linked with schizophrenia. Unfortunately, none of these studies has been able to conclusively identify any specific gene that predisposes to schizophrenia. Typically evidence for linkage is seen on large chromosomal regions, as expected, containing tens or even hundreds of genes. Furthermore, attempts to replicate the findings have rarely been successful leaving a confusion about the existence of predisposing genes for schizophrenia in a particular region of the genome. We have carried out linkage analysis in a set of 62 pedigrees rising from a genetically isolated population of Finland with markers on six chromosomal regions earlier suggested to harbor predisposing genes for schizophrenia, namely 3p, 1 5q, 2,3 6p, 4 8p, 1 20p, 5 and 22q. 6,7 We were not able to find significant evidence for linkage on any of these chromosomal regions. However, some support for linkage was found on all studied chromosomal regions, except 3p.We identified all Finnish patients with schizophrenia born between 1940 and 1969 (n = 29 124). 8 Patients who had received disability pension or free medication because of schizophrenia were identified from two registers of the Social Insurance Institution, and patients who had been hospitalized for schizophrenia were identified from the National Hospital Discharge Register. All three registers were computerized in 1968; ICD-8 diagnostic criteria and codes were used before 1987, since when psychiatric diagnoses have been coded according to ICD-9 applying DSM-III-R diagnostic criteria. We accepted all probands with a 295 diagnosis according to the ICD-8 and ICD-9 numbering scheme; this includes patients with schizophrenia, schizophreniform disorder, and schizoaffective disorder. Parents and siblings were identified from the National Population Register, and information on these relatives was linked to the health care registers to obtain data on their hospital treatments, pensions, and free medications.From these pedigrees, families with at least three affected family members were identified (n = 312), and those with at least one unaffected and three affected children were asked to participate in the study. If both parents were affected, the family was excluded. Diagnoses were confirmed as described in methods. Some families refused to participate and some were excluded because of uncertain diagnoses, and finally 62 families were used for molecular genetic studies out of 102 families who were asked to participate. Three different diagnostic classes were constructed as follows: class 1 contained 137 cases of DSM-IV schizophrenia, class 2 added 23 cases with schizoaffective disorder, and class 3 another 29 individuals with schizophrenia spectrum conditions (schizophreniform disorder, schizotypal personality disorder, schizoid personality disorder, delusional disorder, brief psychotic disorder, and psychotic disorder NOS). Patients with bipolar disorder or major depressive disorder were classifie...