Storage conditions

One objective of seed storage standards is to reduce the frequency at which the most–original sample is regenerated in order to reduce the cost of genebanking and the risk of genetic erosion. Maintaining seed viability is therefore a critical genebanking function.

Appropriate drying procedures are critical to the future longevity of the seed. The appropriate conditions under which seeds should be dried and stored were debated at the recent expert meeting about genebank standards and this is a place where the debate can continue.

What are appropriate relative humidity and temperature ranges? What are the underlying assumptions to predict longevity of seeds?



  1. Fiona Hay said

    Thank you for the opportunity to look at and comment on the draft for the revised standards for genebanks. Focusing on the standards for storage, I have two major issues with the document as it is.

    Firstly, I think there is far too much narration / discussion and a lack of concrete statements. These guidelines should be the definitive source for recommendations that genebanks follow. The text needs to be clear and self-explanatory.

    Secondly, more fundamentally, these revised standards appear to recommend drying at 32% r.h. and 15°C prior to sub-zero storage. This is a considerable change from the 1994 genebank standards, which most seed banks, if they can, follow. The text of those standards includes (for base collections)
    “Preferred: -18°C or cooler with 3-7% seed moisture content (depending upon species)”
    “Drying at 10-25°C and 10-15% relative humidity (r.h.) using either a desiccant or drying chamber is preferred”
    Because seed oil content varies between species, drying according to a single environment means that seeds from different species will have different moisture contents (% fresh weight), hence an acceptable moisture content range of 3-7% is given.

    Have we any evidence that these conditions are reducing longevity? Have any genebanks reported unacceptable rates of viability loss? These existing recommendations continue to be supported by published data for a wide range of species, including genebank data (e.g. Pérez-García et al., 2009) and storage experiments (e.g. Ellis and Hong, 2006).

    In contrast, the proposition that drying to equilibrium with 10-15% r.h. at 10-25°C may reduce subsequent longevity at -20°C arose out of thermodynamic theory, analysis of moisture sorption isotherms, and some experimental storage data on relatively few species [lettuce, sunflower, pea, peanut, and soybean (Vertucci and Roos, 1990, 1993; Vertucci et al., 1994)].

    Different conclusions might have been drawn by different laboratories, in part because of different experimental approaches, as has been discussed in a number of papers. Most fundamentally, Ellis and co-workers have consistently used hermetic storage whilst Vertucci et al. used an open system. That this might have a profound effect on ageing rates was more recently demonstrated by Ellis and Hong (2007); longevity at low moisture contents was significantly reduced in open experimental storage compared with hermetic storage. Most seed banks use hermetic storage. Furthermore, this also emphasizes the potential importance of gaseous environment and a need to make sure the seed-to-air ratio is high or perhaps consider vacuum packing, as used be some genebanks. This is not mentioned in the draft revised standards. The shape of moisture sorption isotherms used to estimate moisture contents at different temperatures also varies.

    What might the impact be if we do move to the draft recommendations being considered?

    Using viability equation (VE) predictions in Kew SID, if we dry rice seeds at 15% r.h. and 15°C using the estimated oil content from Eckey (1954) of 2.2%, we end up with a predicted moisture content of 6.1% fresh weight. The VE prediction of the time for viability to fall by 1 probit at -20°C is 729 years. If rice seeds are dried to 32% r.h. at 15°C they have an equilibrium moisture content of 9.1% and the time for viability to fall by 1 probit at -20°C is predicted to be 103 years.

    It should also be acknowledged that many genebanks use their dry-room as a holding room before processing for storage. In the case of rice, the VE prediction of the time for viability to fall by 1 probit if left in a dry room running at 15% r.h. and 15°C is 62 years; at 32% r.h. and 15°C, the predicted time is 9 years.

    Further, if seeds are to be dried to different moisture contents for storage in the base or active collections (whilst not explicit in the draft, that is one of the implications I think), how is that to be achieved?

    The reason why I raise this scientific discussion is the practical implications of the draft revised standards. As the holder of one of the largest crop germplasm collections, should IRRI risk adopting new genebank standards which haven’t been verified (as far as possible given that we are talking about long term storage) as better than the current conditions in place? What would be the downstream impact on rates of regeneration given perhaps a seven-fold increase in the rate of viability loss?

    I sincerely hope that there can be open discussion on the relative merits of storing seeds at higher moisture contents before revised standards are published. I, together with my colleagues at IRRI, am ready to engage in this discussion.

  2. Robin Probert said

    I share Fiona’s concerns.
    I am also concerned about focusing on the RH at the storage temperature rather than the RH of the drying environment because it will be confusing for the majority of genebank managers. The moisture condition of seeds at the end of drying can be determined with certainty whereas the moisture condition of seeds under long term storage at -18 can only be estimated from models unless very sophisticated instrumentation is used. The draft standard advocates the use of the Kew viability tool to estimate the eMC at the storage temperature and 15% RH and then to use the estimated eMC to derive drying conditions that could achieve 15% RH at the storage temperature. I can see the logic but it relies too much on extrapolation.
    I think it makes much more sense to specify (as before) a drying standard and as Fiona has argued it is difficult to see on the balance of evidence why we should move away from the existing standard. That said, I accept that there is an element of doubt about the optimum moisture content (or eRH) for long-term storage at -18, and I would be comfortable about widening the range of condtions recommended for drying to include the conditions that have been adopted by the US genebank network. Some may see this as a fudge or compromise but it is a way forward.

  3. Christina Walters said

    The reactions of Drs. Hay, Probert and Ellis suggest that we are not close to consensus. But we really are. My colleagues appear to disagree with the RH to which seeds should be dried. The drying RH is hardly debatable once the storage RH are articulated. So, let’s put the potential disagreement on the drying RH to one side for just a moment and focus on the RH to which seeds should be stored.

    Do we concur that the moisture level at which seed longevity is maximum occurs near 15 + 3% RH for seeds of most agricultural species?

  4. christina walters said

    It looks like we are in general consensus that the RH standard for seed storage should be 15 + 3% RH. This RH range is traceable to the published literature as a moisture level giving maximum longevity. It is consistent with research findings in diverse labs using diverse approaches. Uncertainty about aging rates during storage below this RH range was not considered as an element in this particular standard. There is no need to “compromise” on this standard, as has been proposed in the blog discussion, because we are already in consensus about it.

    The temperature standard for long term storage is the freezer, which should read as -20 + 5oC (to be consistent with the format of the RH standard). The basis for this selected temperature range is convenience rather than scientific, as it is the coldest one can get using a single stage compressor. Coincidentally, this temperature range is also the limit of inference for the Viability Equation that uses a quadratic term for temperature response (V.E.) (this form of the equation is used on the Kew SID website). Below this temperature range, the V.E. gives unreliably low estimates of seed longevity.

    My colleagues comments on the blog suggested that the above moisture standard was inconsistent with some of Dr. Ellis’ published findings. I do not see the inconsistency. It would be helpful to resolving the disagreement if the putative discrepancies were pointed out with greater specificity.

    The instructions to the meeting participants were to develop “minimum” standards. Clearly, this allows a genebank to exceed standards if they think it will benefit operations. Hence, if research shows lower storage temperatures will benefit the USDA collection, then NCGRP is free to opt for -80oC freezers or cryogenic storage at < -135oC. Similarly, there is nothing stopping IRRI from storing seeds at RH < 15% if research shows this is beneficial. It is hoped that research findings that compel decisions to exceed storage standards are published so that the genebanking community has that information when it is time for the next round of revisions to storage standards.

    While we are in consensus that long term storage standards are about 15% RH for moisture and -20oC for temperature, we need to be clear to the genebanking community that these standards are based on predictions of seed longevity. There are no actual seed storage data out there to confirm that these conditions will provide the desired longevity of about 100 years. Routine storage in the freezer was only implemented in the 1970s and moisture conditioning using RH was only instituted in the 1990s. We have a long way to go before we can validate our predictions.

    The validity of predictions rests on how well they reflect reality, not how optimistic they are. Dr. Hay provides some interesting predictions of longevity for rice seeds based on the V.E. and Dr. Probert provided a similar analysis for lettuce and barley seeds during the Rome meeting. Unfortunately, these predictions appear to be overly optimistic, likely because they are based on calculations of moisture parameters that are beyond the limit of inference of the V.E. Extrapolating models beyond their limit of inference can give some pretty unexpected results. For example, if I used a similar analysis to predict longevity during cryogenic storage, I would learn that most seeds should be dead in a few months, which obviously contradicts our actual experience. Predictions of longevity in seeds stored at very low RH are unreliably optimistic because the V.E. model continues to calculate an exponential relationship of longevity with water content beyond the low moisture limit of the model’s predictive capacity. The only way to get realistic longevity estimates using the V.E. is to consciously avoid plugging in parameters that exceed the limits of inference for these models. The fact that longevity calculations using a constrained V.E. model are comparable to estimates using other approaches (Appendix, Table 1) should provide confidence in our ability to predict longevity. Comparable estimates using different models do not validate the models, per se, but it leans in that direction. Moreover, the longevity estimates are much more conservative using V.E. constrained by the low-moisture-limit, which provides a more useful management tool when scheduling viability monitoring tests and regenerations.

    Mostly, the limits of inference for the V.E. have been empirically determined (in Dr. Ellis’ lab, my lab and others) at higher storage temperatures and “rules” to extrapolate these to lower storage temperature are tacit. The limit of inference for the V.E. is, by definition, the highest water content giving maximum longevity at the given storage temperature. Two mutually exclusive rules are considered by the revised standards and by these blog discussions: either critical water content is constant with temperature or critical RH is constant with temperature. The former rule (constant critical water content with temperature) is implicit for a standard that uses a single drying treatment (such as drying at 15oC and 15% RH) for all storage temperatures. The assumption of constant critical water content does not hold up to biophysical considerations or empirical evidence. The latter rule (constant critical RH with temperature) is more supportable from a thermodynamic perspective and existing data do not refute it. Hence this is the assumption upon which the revised storage standards are based, and represents an important departure from the 1994 Genebank Standards. The standard of a RH range of 15 + 3% accommodates some degree of experimental uncertainty and differences among species and temperatures that have been detected so far. Someday we may wish to revise this assumption to a constant water potential or plasticization effect.

    A direct consequence of the assumption that critical RH is constant with temperature is that critical water content increases with decreasing temperature. A direct consequence of the alternate assumption, that critical water content is constant with temperature, is that critical RH decreases with decreasing temperature, and this consequence is extremely difficult to rationalize scientifically. These consequences can be easily deduced using the Kew SID water sorption isotherm modules. If you are in doubt, try them – the numbers don’t lie.

    The assumption of constant or changing critical water contents with storage temperature appears to be the underlying sticking point in the disagreement, and it is an issue that cannot be resolved by compromise (even if that were the proper way to resolve a scientific controversy) because the two alternative assumptions are mutually exclusive. In other words, it is a contradiction to develop a storage standard based on RH and then prescribe a single drying regimen that works for all storage temperatures. The narration in the revised standards document tries to illustrate how genebanking procedures must move towards internally consistent standards, and I believe this explanation is warranted.

    After we resolve the issue of target moisture level during storage, then the drying protocols are easy to figure out using water sorption isotherm relationships. Thanks to the Kew SID, reasonable approximations of these relationships are readily available and so the breadth of appropriate drying scenarios can be worked out on paper. In blog responses, people objected to the relatively high RH that could be used to achieve target moistures for -20oC storage. Try the calculations for yourself: it may seem counter-intuitive, but unless the isotherms are in error, the answers are reliable, repeatable and scientifically defensible.

    On a different note, Dr. Hay suggested inclusion of a recommendation for hermetic storage which, of course, is necessary after all the effort of achieving the appropriate moisture level. The text in the technical aspects about using moisture proof containers with general specifications for appropriate materials to achieve desired level of moisture impermeability suffices and addresses many of Dr. Gomez-Campo’s past criticisms without going into too much detail. I would resist the use of the term “hermetic storage” because of its common use by seed groups to describe PVC bags, such as the IRRI-developed “super bag.” These containers are intended for short-term bulk storage and only provide a slight barrier to moisture. Perhaps the term “moisture-proof” used in the document is no better. It’s a balance between too much technical information or over-simplification to the point of incorrectness. Though I don’t feel very strongly about it, I don’t agree with adding a section on vacuum sealing because I do not know of any citable, compelling data or models that this treatment is beneficial. If this data exists, then a section can be added in the technical aspects, but I do not think this should be a standard.

  5. Ehsan Dulloo said

    Keep the discussion going. I hope we would be able to reach some conclusion at some point, but do realise that it may require more evidences and this calls for additional research on the subject.

  6. Ehsan Dulloo said

    To summarize a bit. Currently two standards have been proposed for storage conditions. These relate to the temperature of the storage room and the relative humidity. The standards at the moment read as follows:
    1. Most-original-samples and security back-up samples should be stored under long-term conditions at a temperature less than -15°C. Working samples may be stored under medium term (refrigerator) or short-term (ambient conditions).
    2. All seed collections are stored at an equilibrium relative humidity of 15% ± 3%.

    The rational for the temperature is that new standard should relate to the objective of the storage, i.e whether it is being conserved for the long term or for more medium or shorter usage. There seems to be general consensus that the temperature should be less than -15C for long term storage, although as it is worded, there is no limit to how cold the temperature should be. Christina has suggested in this blog to it should be reworded to -20 +/- 5 C. It has also been said that for all practical reasons this means storage in freezers (-20/-18 C) like the old standard. In addition this standard also clarify that sample conserved for medium term can be conserved in normal refrigerators
    The second standard is a bit more tricky. The change from the old standard is that instead of putting emphasis on the seed moisture content the standard now focuses on the relative humidity of the storage environment that will in turn determine what the SMC would be. Again there seems to be some agreement that 15 +/- 3% eRH is acceptable. The question is: Does this relate to the drying of the seeds? Then what about the RH of the storage room? It seems to me that there are 2 situations that need to be defined. What should be the RH of the storage room in case of open air storage (which is practiced in very few genebanks to my knowledge) and what is the acceptable RH when seeds are packed in hermetic or moisture proof container? is it still 15 +- 3%RH? The argument is that there is still no guarantee that any containers are really moisture proof.
    I think the question which Fiona has asked about the implication of moving forward with these recommendations is very pertinent? We need to be very careful about this as this may influence how genebanks around the world would change their management? To the better or worse? While we must aim at basing the standard on sound scientific evidence, we also have to be vey pragmatic and help to make genebanking practical. As Christina also said, based on what evidence there is for specific species, genebank curators should opt for storage condition that will bring more benefits.

RSS feed for comments on this post · TrackBack URI

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: