Seeing through the fog
“If you're going to while away the years, it's far better to live them with clear goals and fully alive than in a fog.” from What I Talk About When I Talk About Running (2008) by Haruki Murakami (1949- )
Recently a fog rolled in, as a warm weather system made its way through our icy landscape. The warm front brought with it air laden with water. As this humid air drifted through our frigid vistas, conditions were perfect for a winter’s fog. A dense, chilly, earth-bound cloud enveloped everything – obscuring all but the most prominent features, making distant objects impossible to make out.
With the fog, a simple walk was transformed into a voyage of discovery.
The fog transformed everything into a hypothetical.
Did the land rise or fall beyond the horizon?
Was that a person or a tree in the distance?
As we walked forward, each hypothesis was tested. Mysterious features that lurked ahead came gradually into view, revealing their details, their true nature.
As we made our way toward the crest of the rise, a sole, dark figure stood tall in the distance. As we drew closer, the figure proved itself to be a lone tree, a conifer. Moving closer still, the conifer was revealed to be just one of a group of trees, a small stand. Eventually, at the edge of the stand, the stand’s composition was unveiled. Ultimately, the details of the landmark conifer itself were unmasked, so that the very droplets that had obscured its identity could be seen condensing on its needles.
The view of the conifer, distantly, through the fog, provided an incomplete picture of what was to be found after closer examination. With each step forward, our resolution improved, our hypotheses were confirmed or refuted, and new hypotheses were erected in their place.
Science is like that. As we stand distant from the objects towards which we direct our hypotheses, our picture is incomplete. As we view those objects with greater and greater resolution, our hypotheses stand or fall, and new hypotheses supplant them as we push forward with discovery.
A simple experiment investigating the way that plants respond to scarcity of water – drought – provides an excellent illustration of this point.
As they are literally rooted in one place, plants are unable to move about to search out new sources of water when it becomes scarce. Instead, plants have developed a suite of different strategies to contend with drought conditions.
Most land plants make adjustments to their basic physiology in response to drought. As photosynthesis requires water and carbon dioxide to make sugars, plants will alter the way in which they do photosynthesis when water is limiting. They might decrease the rate of photosynthesis, or they may become more efficient in their use of water.
Plant possess tiny pores on leaf surfaces that enable them to take up carbon dioxide for photosynthesis. These pores, known as stomata, also function like the top of capillary tubes – very tiny straws – that enable water to be drawn up through the capillaries from the roots to the leaves where the water will be used in photosynthesis. As such, stomata prove to be sites where plants can lose water – something clearly undesirable when water is in short supply.
Fortunately, plants have evolved means by which to dynamically open and close the stomata. When water is scarce, plants tend to keep the stomata more closed, to prevent water from escaping.
For example, the plants may invoke special developmental pathways that result in the senescence and dropping of leaves to reduce water loss through leaves. It may also involve development of new roots that forage for water beneath the soil. Alternatively, drought may enhance the activity of specific metabolic pathways within the cell – giving rise to specific metabolites that function to protect the plant from dehydration.
Regardless of the mechanisms that plants deploy to contend with drought, these adjustments are largely underpinned by the activity of genes.
Each gene can be thought of us a program that can be brought to bear to serve a particular purpose within the plant. The suite of programs in its entirety is known as the genome. Each cell in the plant body has a copy of the genome – a collection of tens of thousands of genes – that work within that cell to direct the activities of that cell. For example, stomata are each made up of a pair of cells, known as guard cells, each of which has a copy of the genome. The genome in each guard cell responds to cues provided from elsewhere in the plant to direct programs that can open or close the stomata.
We are able to monitor the extent to which each program, each gene, in the genome is being used under a set of circumstances. We do this by detected the read out from each program. This read out is known as a transcript. By detecting the abundance of transcripts for each gene, we have one measure of how much that particular program is being used. Transcript abundance thereby provides an indication of the role a given gene might be playing to enable cells to complete a particular task – such as closing stomata.
We are now able to detect the transcript abundance for all genes in the genome simultaneously. The collective abundance of transcripts for all genes is known as the transcriptome. By looking at the transcriptome, we gain a picture of how the genome responds to a particular stimulus, like drought. By comparing the “drought transcriptome” with the transcriptome from plants that are not experiencing drought, we can determine which genes are “drought responsive” – likely involved in shaping how the plant makes adjustments to water scarcity.
When transcriptome analysis is applied to plants that have experienced drought, some expected trends emerge. Some genes have higher transcript abundance under drought conditions relative to conditions when the plant is well watered. Unsurprisingly, some genes have lower transcript abundance in response to drought, in comparison to when the plant has water. Quite literally hundreds of genes show this kind of response – with either higher or lower transcript abundance in response to a drought stimulus, relative to well-watered conditions.
As might be expected as well, some genes show no difference in transcript abundance, irrespective of whether the plant had water or not. But this latter class of genes is not exactly what it seems…
Analyses of the sort just described are highly controlled – plants that are as genetically similar as possible are all grown at the same time, under the same conditions, and then one half is water-deprived while the other continues to be watered. By taking physiological measurements, like changes in photosynthesis, the plants lacking water are deemed to be experiencing drought. At this time, the plants experiencing drought are all harvested and their transcriptomes analysed. Simultaneously, the well-watered plants are also harvested and their transcriptomes analysed for comparison purposes.
The problem with this kind of experiment is that it stereotypically involves harvesting of the plants for analysis all at one time point. Perhaps not so surprisingly, this is frequently done at a time of the day that is convenient for researchers – like after their morning coffee, or just after lunch. The problem with that approach is that it can miss out on some important stuff.
It turns out that some genes that register as showing no difference in transcript abundance in response to drought relative to well-water conditions, only show that at particular times of the day. For one variety of poplar tree for example, a group of 57 genes shows no difference in transcript abundance in response to water deficit if the analysis is done from plants harvested at mid-day, when researchers would normally measure transcript abundance. However, these same genes have significant increases in transcript abundance in response to drought if transcript abundance analysis is done on plants harvested at the end of the day. Even more remarkably, if transcript abundance is analysed for these same 57 genes from plants harvested in the middle of night, there is, again, no difference between watered plants and plants experiencing drought. But more remarkably still, if transcript abundance is measured from plants harvested right before dawn, this same group of 57 genes has a decrease in transcript abundance in response to drought if the plants used for the analysis were harvested just before dawn!
This is all to say that one cannot simply view a gene as being non-responsive to a stimulus like drought. It may look like it is non-responsive at some times of the day. But at other times of the day the gene is highly responsive. What’s more, at one time of the day it may respond in a “positive” manner, with higher transcript abundance in response to the stimulus, while at other times of the day, it’s response may be negative, with a lower transcript abundance in response to the stimulus. How the program is deployed is, in brief, time-of-day dependent.
Like all organisms, plants do difference things at different times of the day. Photosynthesis, for example, runs during the day, not at night. Generally plants open their stomata during the day to take up carbon dioxide and draw water up through the capillaries in order for photosynthesis to work. At night, the stomata are generally closed to prevent water loss. Obviously, any response to drought is superimposed upon these natural day and night cycles. Consequently, it should be no surprise at all that the transcriptome is reconfigured over the course of the day and night as it appropriate for given points in time.
What is surprising is that sometimes this basic piece of biology gets glossed over in analyses. Genes are said to be positively responsive, or negatively responsive, or non-responsive based on analysed derived from one time point only. While that may be true for that time point, it is important to include the caveat that this is the response documented at that particular time point. It may different if other time points are included in the analyses. Beyond this, it is worth noting that the “responsiveness” of a given gene can be contingent on not only the species under investigation, but can even vary between related individuals within a species, and, crucially, between cell types within a given individual.
Crucially, we only know these things when we investigate hypotheses using higher resolution, including the capacity to measure many, many possible responses at once. Transcriptome analyses can be done across multiple time points, and involve looking at all genes in the genome, all tens of thousands of them, at once. This level of resolution enables us to see through the fog, to provide a more complete picture of how organisms respond to stimuli.
Unfortunately, the same cannot be said for other analyses.
As but one example, recently, much has been made about the efficacy of Vitamin C supplements to combat the common cold - or, to be more specific, about the lack of efficacy of vitamin C supplements to do just that. Vitamin C supplements are effectively a chemical stimulus, and people have been on the lookout for some sort of desired response induced by that stimulus. To date, there is no strong evidence to support the hypothesis that Vitamin C supplements do anything of the sort for the general population.
But has the hypothesis really been refuted?
There is an argument to be made that the right experiment has not yet been done. Has there been an effort made to ensure that the subjects in such hypothesis tests are in the same physiological state during the experiment? We know, for example, that diet can have a profound effect on the gut microbes of individuals. We also know that gut microbes metabolise our diets in different manners, leading to changes in human physiology. If one doesn’t control for diet, and other factors influencing our overall physiology, include our resident microbes, how is it possible to adjudicate the efficacy of any supplement?
Interestingly, there is evidence to support the hypothesis that physiological state might influence the efficacy of Vitamin C in preventing the common cold. Almost precisely a year ago, Hemilä and Chalker reported that five trials with 598 participants exposed to short periods of intense physical stress (e.g., marathon) had their risk of common cold halved by Vitamin C supplements. This suggests that physiological state may shape the efficacy of Vitamin C in combating the common cold.
Aside from their physiological state, have some subjects in supplement efficacy studies consumed the supplement at one time of the day, and others at a different time of day? When were these times relative to consumption of food or liquids? Finally, what was the measure of a “response” when the supplement was provided? Was it a wide-ranging, relatively unbiased approach like transcriptome analysis, or did it focus on one single measure? When was the measurement taken? Was it only taken once? Might the measurement vary, contingent on time of day for example?
All told, Vitamin C efficacy studies still seem in a fog. They haven’t seemed to take the step forward to have the resolution that is really needed to adequately test the hypothesis of efficacy. The best we can say is that studies testing the efficacy of Vitamin C supplements have been unable to validate or refute hypothetical efficacy. It may be that vitamin supplements have no effect, but the studies haven’t been done to show that one way or another. In all likelihood, vitamin supplements do invoke something – they are, after all, a chemical stimulus – it just remains to be seen if that is the response that some people desire.
Vitamin C supplement studies are just one example of the pitfall that arises from a foggy situation – we can draw conclusions based on an incomplete picture. We frequently find ourselves in situations where the direction ahead is less than clear. Our resolution is not what we need to make a judgement one way or another. The only way forward is to step toward the horizon, make observations at a higher resolution, test them against our hypotheses, and adjust our assumptions if need be – eventually the veil of the fog will be lifted, and our understanding will be more complete.
Disclaimer: This piece is not advocating for or against the use of Vitamin C supplements to combat the common cold. This piece should not be construed as medical advice of any kind. The piece is only using analysis of Vitamin C supplements as a means to combat the common cold as an example of where our knowledge is less than complete, but where statements are being made as if we have complete knowledge about efficacy. It is the opinion of this author that our understanding of the impact of Vitamin C supplements in combating the common cold is far from complete. Any use of vitamin supplements should be done in consultation with a medical health professional.
Images: All photographs by Malcolm M. Campbell.
David LA, Maurice CF, Carmody RN, Gootenberg DB, Button JE, Wolfe BE & Turnbaugh PJ (2013) Diet rapidly and reproducibly alters the human gut microbiome. Nature (early online)
Hamanishi ET, Raj S, Wilkins O, Thomas BR, Mansfield SD, Plant AL & Campbell MM (2010) Intraspecific variation in the Populus balsamifera drought transcriptome. Plant, Cell & Environment 33: 1742-1755
Hamanishi ET, Thomas BR & Campbell MM (2012) Drought induces alterations in the stomatal development program in Populus. Journal of Experimental Botany 63: 4959-4971
Hemilä H & Chalker E (2013) Vitamin C for preventing and treating the common cold. Cochrane Database Syst Rev 1
Louis P & Flint HJ (2013) How our gut microbes influence our behaviour. Journal of Neuroendocrinology 25 517-518
Wilkins O, Bräutigam K & Campbell MM (2010) Time of day shapes Arabidopsis drought transcriptomes. The Plant Journal 63: 715-727
Wilkins O, Waldron L, Nahal H, Provart NJ & Campbell MM (2009) Genotype and time of day shape the Populus drought response. The Plant Journal 60: 703-715