Forum:Bioinformatics For Clinicians - What To Teach Them?

Question

[j]

We are preparing a course in bioinformatics for clinical practitioners, an are in the process of planning the contents of the course. I would appreciate your thoughts on the content, and if you know any existing resources that have been planned for similar purpose (e.g. existing online courses or course descriptions?).

Especially, what do you think are the most relevant things about bioinformatics that should be taught to clinical practitioners during a short course?

Draft of my initial plan include e.g.:

• What bioinformatics is and how you can apply it in clinical research and patient work
• Basic bioinformatics tools (e.g. databases, maybe some simple data analysis methods) and their practical usage
• Theoretical background to some common bioinformatics methods (e.g. machine learning/classification?), how to interpret and understand publications that utilise bioinformatics
• Challenges and pitfalls of bioinformatics
• Using modern high-throughput technologies (sequencing and other -omics technologies) as part of clinical work
• Bioinformatics as part of personalised and translational medicine, case studies
• How to take into account the needs of bioinformatics when planning and executing a clinical study
• How to be a bioinformatics expert, what kind of additional education is needed (e.g. statistics, programming, computer science in general etc?)

The focus group in this particular course are students who are/will be medical doctors, and therefore I believe the topic should be on the practical side, and contain lot of examples from how bioinformatics has been used to solve real life research questions, as well as applied in clinical patient work (if you have good examples of this, they would also be appreciated!).

Answer 1

[Neilfws]

If medical students are anything like they were when I studied with them 20 years ago, they will have an enormous course load. Most will be reluctant and/or unable to add extra study material deemed not directly relevant to the everyday practise of medicine.

So I think they should know:

• that bioinformatics exists
• that it is useful and important
• that bioinformaticians are available and happy to help if required :)

Answer 2

[DG]

I have given some lectures to Pathology residents about bioinformatics (in the context of NGS and clinical applications) as well as bioinformatics oriented talks for other clinicians along similar lines. It all depends on how much time you have, whether this is a typical course or just a few lectures. My advice:

• Keep it light. DO NOT get bogged down in too much algorithmic detail. They don't care and to be honest, they don't need to know
• Give as many examples as possible from case studies and clinical applications
• Do provide some insight in to both what bioinformatics can do and why it is essential, and also where some of the challenges are
• I would talk up the need for bioinformaticians and clinicians to work together in teams. How clinical insight makes our jobs easier, how we can make theirs easier, etc

Answer 3

[Michael Schubert]

pvalues?

[Biomonika (Noolean)]

Thanks for sharing!

Answer 4

[Charles Warden]

I think information on variant characterization will likely be useful

Here is an important NEJM article:

http://www.nejm.org/doi/full/10.1056/NEJMoa1306555

Likewise, some discussion about variant callers (VarScan, GATK, etc.), annotation programs (ANNOVAR, SnpEff, SeattleSNP), etc. will probably be relevant. Importantly, a discussion of the impact of these tools on the results are likely important. Here are some potentially useful resources:

• http://www.ncbi.nlm.nih.gov/pubmed/22424236
• http://www.ncbi.nlm.nih.gov/pubmed/23340846
• http://www.ncbi.nlm.nih.gov/pubmed/24297550

I say this because I suspect exome / genome sequencing will eventually become a common piece of information sought by patients. I think current discussions about 23andMe are probably a decent starting point for the technical discussion. Here is some recommended reading on that topic:

• http://genomebiology.com/2013/14/11/139
• http://www.genomesunzipped.org/2013/12/pertinent-and-non-pertinent-genomic-findings.php
• http://daviddobbs.net/smoothpebbles/feds-muzzle-23andme-decide-public-best-be-ignorant-about-genetics [lots of good links]
• http://daviddobbs.net/smoothpebbles/i-got-your-23andme-v-fda-links-right-here/ [shorter list of links]

[and my own blog posts]

• http://cdwscience.blogspot.com/2011/02/thoughts-on-my-23andme-results.html
• http://cdwscience.blogspot.com/2012/06/my-23andme-results-getting-free-second.html
• http://cdwscience.blogspot.com/2013/11/my-take-on-fda-warning-to-23andme.html
• http://cdwscience.blogspot.com/2013/12/who-was-primary-audience-for-fda.html

Answer 5

[amcrisan]

First I would assess what knowledge of genetics physicians have. It's my understand that only a limited portion of medical school is dedicated to an understand of genetics.

I like to think about bioinformatics and clinicians in the three following groups of concepts

1. Genetics and Medical Decision making (build context):

   • classic examples like ER+, HER2-neu
   • what genetic information we have post-human genome era
   • how bioinformatics emerged to help handle all that data

2. What is a bioinformatician and what do they do?

   • Wrangling of genomic data
   • Looking for variants (list variant types.. and what their impact may be)
   • How do they accomplish this (with the power of statistics, computer science, and domain knowledge combined!!!)??

3. How can a bioinformatician help you?!

   • How to talk a little statistical
   • Where to find one.. or if you don't have one near you, how can you ask one a question?

Answer 6

[Giovanni M Dall'Olio]

I would give them an overview of the most relevant databases, and explain how to use them.

• all the NCBI databases, and entrez
• the OMIM and other databases for genetic diseases
• maybe an overview of the UCSC browser, highlighting some relevant tracks, and describing which information can be accessed through it

Answer 7

[sarahhunter]

You might find these pages from the UK's PHG foundation useful in deciding how to focus your course:

http://www.phgfoundation.org/project/wgs

Recent advances in DNA sequencing technology are making rapid, affordable genome sequencing increasingly viable in health services. This will pave the way for increasingly personalised medicine and better medical care and public health, and offers immediate benefits for patients with rare diseases and cancer. So what will it take to make clinical genome sequencing and analysis a reality?

This project is examining different aspects of clinical genome analysis in greater depth, to explain and address what barriers need to be crossed in order to use genome sequence data for practical, clinical purposes.

Answer 8

[Asaf]

I would like to suggest teaching some very basic protein related tools like structure prediction and comparison. Sometimes MDs have sequences of non-functional proteins, usually they just spot the mutation and report it. They need to known that there are tools that will help them understand what the mutation cause, how it influences the structure or the action of the protein. As I see it, they don't need to be able to use these tools, just to know that they exist.