blog

Harriet Dashnow is a Bioinformatician and PhD candidate at Murdoch Childrens Research Institute, Royal Children’s Hospital, in Melbourne. She has several posters on F1000Research, and we asked her about her experiences with sharing posters, and about her work.

Why do you share your posters on F1000research?

I upload my posters because I want to increase the number of people who can access them. At conferences there are often 100+ posters and people may not have time to see every one.
I also try to put a a link to my poster up on Twitter, or similar before the poster session, so people can see exactly what my poster is about before they come to see it.
Having my poster online also means that people don’t have to take low quality pictures of it using their phone if they want to read the details later, they can just search for it online.
I have uploaded posters to both F1000 and Figshare. The choice of which generally comes down to if the conference recommends one or the other. It’s useful to have all the posters and slides from a given conference in one place.

Have you had any response from people who saw your poster online?

I have had people who saw my poster online (via Twitter) come up to to talk to me at that same poster at a conference.

You uploaded your first poster on the old F1000Posters site, and we’ve obviously made some changes to the site since then. What do you think of the new poster submission system on F1000Research?

Last time I uploaded a poster I was impressed by how quick is was. Great job!

Can you tell us about your most recent poster? What’s it about?

Short tandem repeats (STRs) are short (2-6bp) DNA sequences repeated in tandem. We know that STR variation can cause disease in humans, for example neurological and developmental disorders. STR variation is also important for gene regulation and other cellular processes.

In this poster, I compare software tools to genotype STRs in next-generation sequencing data. I use real and simulated data to look at how well the most popular genotyping tools agree with each other, and to investigate the importance of read mapping. Comparing two of the most commonly used tools (RepeatSeq and LobSTR) they only agreed that a particular STR was variant 14.6% of the time. And of those loci they both called variant, they only agree on the specific genotype about half the time. Compare this with concordance between indel callers, with is usually >40% and it’s clear the STR callers have some room for improvement. Thankfully, this is an active area of research, and there are often releases of new tools and updates to existing ones. So I think STR genotyping will become increasingly common and accurate.

Why is it important to choose the correct algorithm when genotyping STRs?

We may need to make research or clinical decisions based on the variants we call, so it’s important to know how accurate those variant calls are. For short tandem repeats, most of the algorithms are very new and hadn’t been systematically compared yet. In this poster, I try to address some of these issues by comparing the most commonly used software tools.

Will you be publishing that work as an article soon?

Still a bit of work to go before it gets to that stage, but hopefully soon.