Scientists have identified a series of human sperm biomarkers that can accurately predict the chances of the offspring having autism spectrum disorder ( ASD) – a potentially important finding for better understanding and managing the condition.

Although the study involved a small sample size of only 26 fathers, the strong correlation between the biomarkers and ASD suggest that this thread of research could be worth pulling on to help explain some of the mysteries that still surround the development of autism.

In particular, the researchers looked at sperm epigenetics: changes in the way that genes are expressed and processed by the body, rather than alterations in the underlying DNA code, and often linked to environmental factors.

"We can now potentially use this to assess whether a man is going to pass autism on to his children," says biologist Michael Skinner, from Washington State University.

"It is also a major step toward identifying what factors might promote autism."

The study included samples from 13 men who had fathered autistic children (case) and 13 men who hadn't (control). In an analysis of DNA methylation regions – particular chemical modifications to gene expression – the team identified 805 potential biomarkers.

They then blinded eight of the samples and reanalysed them to determine whether these were case or control. In a further blind test of 10 additional sperm samples, the scientists were able to use their DNA methylation findings to predict whether or not the men had fathered autistic children with an accuracy rate of around 90 percent.

Not only could the study be expanded to predict the chances of men fathering autism, but it could also provide clues as to how these epigenetic changes come about in the first place, according to the researchers – perhaps giving us a way to do something about them.

"We found out years ago that environmental factors can alter the germline, the sperm or the egg, epigenetics," says Skinner. "With this tool, we could do larger population-based studies to see what kinds of environmental factors may induce these types with epigenetic changes."

With a sharp rise in children diagnosed with ASD across the last few decades, experts are keen to learn as much as possible about the causes of the condition – and if we're able to predict it, then we should be able to manage it better too.

Although the heritability of autism isn't something scientists fully understand, fathers are more often linked to passing on ASD than mothers, making the findings of the current research even more valuable.

While the small sample size of this study means it's too early to start making broad generalisations, the research does show that this is a link worth pursuing – and a bigger clinical trial following the same approach is already being planned.

"Although a large clinical trial is needed to further validate the biomarkers and potential diagnostic, the current study provides the proof of concept for the assay and biomarkers," write the researchers in their paper.

The research has been published in Clinical Epigenetics.