Background Attention Deficit/Hyperactivity Disorder (ADHD) is characterized with inattention, hyperactivity and impulsivity, while Autism Spectrum Disorder (ASD) is characterized with impaired social communication and interaction, restricted and repetitive patterns of behavior and activities. Phenotypically, both appear to be quite different. However, clinically it has been found that 20–50% ADHD children meet diagnostic criteria for ASD, while 30–80% ASD children meet diagnostic criteria for ADHD. These rates of comorbidity between the two disorders imply that they may share underlying etiologies. Although recent evidence has suggested a genetic overlap between ADHD and ASD, little progress has been made so far to identify the specific genes involved. The shank synaptic scaffold proteins, which are encoded by SHANK genes and located at the post-synaptic density of glutamatergic synapses, have been reported to be associated with a number of psychiatric disorders, including ADHD and ASD, separately. The aim of the current study is to investigate whether there is a genetic overlap between ADHD and ASD in SHANK genes. Methods In the present study, only male children with Chinese ethnicity, aged between 6 to 12 years, were included. 298 ADHD family trios (including ADHD children without ASD and their biological fathers and mothers), 134 ASD children without ADHD, and 109 children with both ASD and ADHD were recruited from public hospitals in Hong Kong. 232 normal control children were recruited from local primary schools. A number of single-nucleotide polymorphisms (SNPs) from SHANK2 and SHANK3 were selected for genotyping. After quality control, 14 SNPs from SHANK2 and 7 SNPs from SHANK3 were included for data analyses. For ADHD family trios’ data, we constructed ADHD cases and matched pseudo-control, which were based on transmitted and un-transmitted parental alleles using the Haplotype-Relative-Risk (HRR) principle. The pseudo-control supplemented the normal control to increase the statistical power. Likewise, ASD children with or without ADHD were also paired with samples of normal control and pseudo-control for association analysis. PLINK and KGG were used for statistical association analyses at SNP-level and gene-level, respectively. Results When association analysis was done between ADHD children without ASD and samples of normal control and pseudo-control, the T allele of rs7106631 of SHANK2 significantly decreased the ADHD risk (p=0.001, OR=0.70). When allele frequencies’ comparisons were conducted between ASD children with or without ADHD and samples of normal control and pseudo-control, the same protective effect of rs7106631 T allele was identified with OR=0.74 (p=0.011). Further association analysis conducted between all clinical children (i.e., ADHD children without ASD & ASD children with or without ADHD) and samples of normal control and pseudo-control yielded more significant results for rs7106631 (p=0.0003, OR=0.72). Additionally, significant associations were found for other SHANK2 SNPs, namely, rs7113016 (p=0.027), rs1073294 (p=0.008), rs11236616 (p=0.003), rs10899158 (p=0.048), and rs9888288 (p=0.004). However, no significant result was detected for SNPs of SHANK3. Subsequent gene-level analysis based on whole dataset suggested the significant role of SHANK2 in both ADHD and ASD (gene-level p=0.003). Discussion Consistent with previous findings in Western populations, this current study provides additional evidence for the significant association of SHANK2 with ASD from a Chinese population. A new finding rarely reported in previous studies is the association of SHANK2 with ADHD. This implies a shared genetic susceptibility between ADHD and ASD. Both disorders do share a number of characteristics, including male preponderance, increased level of oxidative stress, and abnormal neuronal connectivity, reflecting potentially common underlying neurodevelopmental abnormality. Speculatively, the identification of SHANK2 from this current candidate gene study as associating to both ADHD and ASD may provide the common genetic etiology to explain their shared characteristics and higher-than-chance comorbidity. Further studies examining the functional abnormalities of the identified gene, SHANK2, are warranted in the future.