New research has boosted the number of genes definitely known to play a role in autism from nine to 33.
More than 70 other genes were pinpointed that are also likely to be linked to the developmental disorder, which impairs a person's ability to socialise and communicate.
Small variations in up to 1,000 genes may eventually be found to contribute to the condition, the evidence suggests.
The newly-identified genes are believed to be important to critical brain processes affecting the formation of nerve networks and neural connections.
Dr Kathryn Roeder, one of the researchers from Carnegie Mellon University in Pittsburgh, US, said: "This makes sense because typical development of brain cells require intricate co-ordination among thousands of genes and appropriate communication between cells to ensure development of the brain, the most complicated organ in the human body."
The findings support a theory that says synchronisation between front and rear brain areas is lower in people with autism.
The scientists, whose work is reported in the journal Nature, analysed more than 14,000 DNA samples from autistic children, their parents, and unrelated individuals.
Statistical tools were developed that enabled the researchers to assess the impact of both inherited gene mutations and those that spring up spontaneously in sperm and eggs.
The research expanded the number of known definitive autism genes almost four-fold, from nine to 33.
In total, small, rare variations in 107 genes were thought to confer a relatively large increase in someone's chances of being on the autistic spectrum, which encompasses the full range of mild to severe effects of autism.
The scientists predicted that small differences in about 1,000 genes will eventually be shown to increase autism risk.
More than 5pc of the autistic individuals studied had new, non-inherited, loss-of-function gene mutations.
Co-author Dr Bernie Devlin, from the University of Pittsburgh, US, said: "I am confident that the list of autism genes will expand rapidly because there are already many more samples sequenced.
"What goes awry is a harder question, but the ever increasing list of genes involved will surely provide pieces that could solve the puzzle of autism."