 |
| Lots of leads? @ Wikipedia |
- the microbiome (microbiomics - the study of bacteria, or more specifically, the bacteria that inhabit us),
- the epigenome (epigenomics - the study of the various ways in which gene function can be modified or altered),
- and the metabolome (metabolomics - the study of the low molecular weight metabolites that we all produce [and excrete] on a daily basis).
If I had to 'pick an omic' (now there's a new saying) that I perhaps have had most experience with, it would have to be metabolomics and the whole biomarker thing based on some of the technology and work I've been involved with (albeit not as lead). Indeed metabolomics, whilst sounding very, very complicated is actually quite a simple science in terms of its basic tenets: chemically separate sample, chemically analyse sample, statistically analyse results, organise statistical results. It obviously gets a little more complicated as you delve further into the technology (mass spectrometry, nuclear magnetic resonance) and the statistical methods (principal components analysis) involved, culminating in literally volumes of data.
Another layer to the whole metabolomics (and other -omics) field is the area of systems biology (and bioinformatics). Although quite a nebulous term, systems biology can involve back-tracking through your metabolites of interest in order to both look at interactions between the metabolites and eventually from which systems they derive and whether there is a connection between those systems. It's data organisation for biology. A good example of the process was demonstrated in a recent post on carnitine and autism (bearing in mind this was more of a proteomics kinda thing).
'Get to the point' I hear you cry, as I bring to your attention a paper by Randolph-Gips & Srinivasan* (open-access) discussing the modelling of autism with a systems biology slant. The first thing to say about this paper is that whilst no new data is actually presented, it represents a pretty good review of the complexity of autism based on quite a lot of the more biochemical research undertaken. So for example the key areas of antioxidant, gastrointestinal, mitochondrial, immunological and neurological functions are included alongside a whopping 183 references; someone has done an awful lot of background reading for this paper. The value-added bit to the paper (IMHO) is the suggestion that modelling the variety of presentations in these various areas in cases of autism might actually be quite a useful thing to do.
I wouldn't claim to be an expert on systems biology but do recognise some of the suggestions mentioned in this paper. So things like hierarchical modelling and identifying subgroups (yes, endophenotypes again). Also taking into consideration that autism is not a static entity (the authors refer to autism as a 'disease' but I would disagree with that description) and the introduction of things like dynamic time warping to take maturational changes for example, into account.
I know some people won't be convinced by papers like this. They'll either question the various research included in the review and its relevance to autism, or point out the volumes of research missing from the review. I agree that this is a complex area; autism research, the thousands and thousands of findings reported on the autisms over the years, is complicated. The point of papers like this however is how we go about organising all that knowledge into something manageable, responsive to the addition of new data and useful in terms of discerning the various connections between often disparate areas of research onward to hypothesis generation. Realising also that although research findings may be presented with autism in mind, does not mean that they aren't also relevant to other conditions (and indeed possible comorbidities) is an important part of this systems biology approach.
Now all we need is someone to put some money and time into building that massive database or indeed incorporating new knowledge into existing databases. Any takers?
----------
Randolph-Gips MM, Srinivasan P. Modeling autism: a systems biology approach. J Clin Bioinforma. 2012; 2: 17.
----------
'Get to the point' I hear you cry, as I bring to your attention a paper by Randolph-Gips & Srinivasan* (open-access) discussing the modelling of autism with a systems biology slant. The first thing to say about this paper is that whilst no new data is actually presented, it represents a pretty good review of the complexity of autism based on quite a lot of the more biochemical research undertaken. So for example the key areas of antioxidant, gastrointestinal, mitochondrial, immunological and neurological functions are included alongside a whopping 183 references; someone has done an awful lot of background reading for this paper. The value-added bit to the paper (IMHO) is the suggestion that modelling the variety of presentations in these various areas in cases of autism might actually be quite a useful thing to do.
I wouldn't claim to be an expert on systems biology but do recognise some of the suggestions mentioned in this paper. So things like hierarchical modelling and identifying subgroups (yes, endophenotypes again). Also taking into consideration that autism is not a static entity (the authors refer to autism as a 'disease' but I would disagree with that description) and the introduction of things like dynamic time warping to take maturational changes for example, into account.
I know some people won't be convinced by papers like this. They'll either question the various research included in the review and its relevance to autism, or point out the volumes of research missing from the review. I agree that this is a complex area; autism research, the thousands and thousands of findings reported on the autisms over the years, is complicated. The point of papers like this however is how we go about organising all that knowledge into something manageable, responsive to the addition of new data and useful in terms of discerning the various connections between often disparate areas of research onward to hypothesis generation. Realising also that although research findings may be presented with autism in mind, does not mean that they aren't also relevant to other conditions (and indeed possible comorbidities) is an important part of this systems biology approach.
Now all we need is someone to put some money and time into building that massive database or indeed incorporating new knowledge into existing databases. Any takers?
----------
Randolph-Gips MM, Srinivasan P. Modeling autism: a systems biology approach. J Clin Bioinforma. 2012; 2: 17.
----------
Randolph-Gips MM, & Srinivasan P (2012). Modeling autism: a systems biology approach. Journal of clinical bioinformatics, 2 (1) PMID: 23043674
Now you're just being silly. Actually looking at the many (many, many, many) biological disruptions in people with autism, realizing that a disruption in one area can cause problems in other areas, and taking a more systems approach to tackling the problems? That's crazy talk.
ReplyDeleteDidn't anyone ever tell you that all of the biological disruptions seen in autism are co-morbid conditions that have absolutely nothing to do with the core of autism?
But seriously, a paper like this is a good thing. There are some medical practitioners who are already on board (and have been for some time) with the idea that treating the biological disruptions in autism can lead to improvements in the core symptoms of autism. It is long past the time for the rest of the medical community to do the same.
Thanks as always MJ.
ReplyDeleteI know, I know, crazy talk indeed isn't it that someone might actually want to try and 'join the dots' in a coherent and systematic manner.
I think we are going to see more and more of these kinds of papers start to emerge as time goes on. Machine learning and the whole bioinformatics arena are upcoming areas in lots of aspects of medicine; its only a matter of time....