There are many tests of inheritance based upon sibling information for diseases that have late onset. The N-test is one these tests, which utilizes information from affected siblings. The N-test excludes recombinant families from the analysis, when applied to haplotypes. This study proposes a non-parametric statistical test (Allele-based N-test) for the detection of genetic linkage in general pedigrees, investigates properties of this test statistic such as significance thresholds and power, and finally compares the proposed test with two other standard widely-used non-parametric tests (Sall and Spairs). The test statistic is motivated by the need of a statistic that can use both recombinant and non-recombinant affected offspring for mapping linkage to highly polymorphic haplotypes such as those in HLA region. This research also studies the effectiveness of CIDR's (Center for Inherited Disease Research, John Hopkins University) data cleaning rules for Mendelian inconsistencies and proposes alternative rules that can potentially perform better.