EEB240 Lab 5
Toxicity Bioassay Part II: Putting it all together
Where to meet: RW124
To do before the lab:
Read this lab and its steps in their entirety. Make sure you understand what you are to do before coming to class!
What you’ll need for this lab:
As for Lab 1, you will need a device that can open and allow you to manipulate an Excel spreadsheet and that has R and R Studio. Please let us know if you need a device.
Background
Experiments are carefully designed to investigate specific questions, and they ultimately help scientists understand the world. It is easy to imagine that, once an experiment is completed, the hard work is done. In fact, once the experiment is complete, the hard work is just beginning! After an experiment, scientists begin that painstaking task of assembling, scrutinizing, and finally analyzing the data. Our ability to make inference from our analysis is only as good as the data, so we have to objectively ask whether the data we collected during the experiment is reliable. For example, were there any procedural flaws (e.g., accidents, or known inaccuracies) that occurred during the experiment that might affect our results, or our ability to make inference? If so, is there anything we can do at this point to mitigate those problems? An example might be excluding a sample with a concentration that is known to be incorrect. Or, we can ask: even if there were not any known procedural flaws, is there any signature in the data that suggests we made an error during our experiment? For example, if survival is greater than 100% in one of our treatments, then it’s likely that Daphnia were born in one of our jars and we counted the babies. More generally, after any experiment, it is imperative that scientists objectively explore the data so that we can be confident in the conclusions drawn during our final analysis.
This brings us to a key component of scientific investigation, which is that the results must be repeatable. In other words, if another scientist wanted to independently verify your results, they could exactly reconstruct your experiment and achieve the same (or very similar) result. In order to ensure that the result is repeatable, the scientist must describe, in writing, the exact steps they used to both set up the experiment, as well as any decisions they subsequently made while handling the experimental data. These data-handling decisions must be transparent, and they should be defensible, rather than arbitrary.
Our goal during this lab is to explore the data we collected during the LC50 experiment, to think about what the data are telling us, and to report what decisions we made when handling the data. We will also contextualize these data into what we learned over the series of labs in the course.
Terminology:
A treatment in this experiment represents a group of jars with the same concentration of NaCl. For example, we have a 500 mg/L NaCl treatment, and a 7000 mg/L NaCl treatment.
In our experiment, each treatment is replicated twice. A replicate is a jar that represents a treatment, and we have two replicates per treatment. For example, we have two jars with the 500 mg/L NaCl treatment, and each one of these jars is termed a replicate of that treatment.
A control is a special kind of treatment where the relevant manipulation (in this case, NaCl addition) is not performed. A control allows us to gauge what response occurs in the absence of the variable that was manipulated (in this case, NaCl concentration was manipulated).
Procedure
1) Chronic Toxicity Bioassays:
Last week, your group exposed Daphnia to various treatments of NaCl (0 – 7000 mg/L) both in hard water and soft water. This week, we will combine the data from all 6 groups in your lab section to build and explore Dose Response Curves. Each group in the lab has contributed to this lab’s dataset on LC50, which is the dataset that you will all use in class and for your report. Then, we will use what we know about salinity concentrations in the GTA from Lab 1 and Lab 3 to assess whether our urban activities are making aquatic habitats toxic to organisms such as Daphnia.
Step 1: Scrutinizing Your Data
1. Your instructor will have assembled all the data from your group in an Excel sheet; it will be available via Quercus. Download your lab’s data (it will be “yourPRAsection Lab Data.xls”; do NOT download the alternative Excel file “yourPRAsection R Analysis Data.csv”), and open it in Excel. Your instructor will go over the structure of the data with you.
2. Let’s begin by assessing overall data quality. First, is there mortality in the controls that is cause for concern? If there is high mortality (equal to or greater than 20%, on average) in the controls, then you should keep this in mind in the final interpretation of your result.
Answer the following question below: what was the mean survival in the two control treatments, do you feel this level of mortality is cause for concern with reference to the 80% threshold, and how might you explain any mortality (be it high or low) in your controls (2 pts)?
3. Next, have a good look at the conductivity data. Each treatment is replicated in hard water, and in soft water, and replication was carried out by independent groups of students. If the appropriate amount of NaCl was added, and if our conductivity meters were working correctly, then the conductivity values should be similar in the same treatments (e.g., both 1000 mg/L NaCl treatments in hard water should have similar conductivity values). Are there any treatments that have unusual conductivity readings? Here you will need to use your judgement, as every good scientist must do: you must decide if you have any unusual data, and make a decision about what to do with unusual data. There is no correct answer, we just want you to think about the data, the sources of experimental error, and make a judgement call.
Here are some options you can use to help assess whether you have any unusual datapoints:
i) By just looking at the data, do you see any conductivity readings that seems much too high or low?
ii) Or try plotting NaCl treatment vs conductivity in Excel (scatterplot), and see if the conductivity in any replicate is very different from the other conductivity readings.
iii) Or perhaps make an objective benchmark: for example, an usual conductivity reading could be any replicate that was more than, say, 1.5x the conductivity you expect that replicate to be based on other replicates in that treatment,
iv) Any other approach you believe is sensible and justifiable.
In the space below, describe any assumptions you make (e.g., ‘no pair of treatment replicates differed by X µs/cm, so I believe the correct amount of NaCl was added to each replicate’), as well as any alterations you make to the data (e.g., remove a replicate because you are not confident the appropriate amount of NaCl was added, based on the conductivity readings). Describe and explain why you made any decisions. (2 pts)
4. Recall that Cl- was estimated simply as NaCl (mg/L) * 0.607, given that NaCl is 60.7% Cl- by weight. Look at your Cl- estimates in your Excel sheet. Each concentration of NaCl should correspond to the same Cl- estimate. If they don’t, then think about re-calculating Cl- estimates before proceeding. You do not need to report that you did this, as the multiplication factor is not controversial or debatable.
5. Have a final good look at your data. Is there any other cause for concern? For example, was there a replicate where a mother clearly gave birth? If so, how will you deal with this replicate (exclude it, remove the babies from the final count, other)? Describe any data handling and manipulating that you did (that is not already described above) in the space below. If you did not make any other change, please say that there were no other changes (1 pt).
Step 2: Graphing in Excel
6. Visualize how you would draw the dose-response curve for Cl-. What should go on the y-axis, what should go on the x-axis? Plot the 24 hr survival data for the soft water treatment using Excel. Do not include the replicate with water from the field. Make sure you label your axes and include any relevant units. For now, just visualize the plot on your computer; in a subsequent question you will be pasting this plot into this document.
7. Still looking at the graph, is there any survival datapoint that doesn’t fit the data well? If it doesn’t, what about that data point makes you concerned? Do you know if something went wrong during the experiment (i.e., is there a known problem with this treatment?). If there is, report this to your TA and discuss it. This will not affect your mark, but it may affect the conclusions of your study, which is why it should be discussed.
8. Next, plot the 48 hr mortality data for soft water on the same figure (not including replicates with field water). Give this graph a title in Excel: “Daphnia survival in response to Cl- concentrations in soft water at 24 hrs and 48 hrs”. Are there any survival datapoints that really don’t fit the data well, as in (7) above? If there are, report this to your TA and discuss it. In the space below, report any new decision you made (e.g., exclude a replicate not already excluded in the previous steps) after visualizing the data, and even if you made no data manipulation/exclusion at this step, make sure you report this (1 pt). When you are satisfied with your figure, paste it below, below any text you provided that reports on your handling of the data (4 pts – for the graph).
9. Finally, plot the same two dose response curves for the hard water treatments. As above, if there is a problem with your data, report this to your TA and discuss it. This will not affect your mark. In the space below, report any decision you made (e.g., exclude a treatment) after visualizing the data and, if there was an issue, discussing it with your TA (1pt). When you are satisfied with your figure, paste it below, below any text you provide to describe the handling of the data (4 pts– for the graphs).
10. Now you should be satisfied that the data are as reliable as they will get, and you have reported (in the steps above) how you handled the data.
11. There is no need to quantify the LC50 yet, but just by eye, does the LC50 look different in hard water vs soft water? Does the LC50 change over time? Think about whether any difference you see makes sense; if you don’t see a difference, does that make sense? Recall that your goal here is to understand your data, so if the data do not make sense to you, ask your TA for help.
Step 3: Analyzing your lab data in R
12. Save your LC50 data, along with any changes you made. Next, you need to download your lab section’s data from Quercus; it will be entitled “yourPRAsection R Analysis Data.csv”. First, any changes you made in your Excel document in steps 1 – 11 also need to be made in your .csv file. In other words, if you altered the data in any way in Excel, you need to change it in your csv file too; refer to the notes you made during the previous steps to do know which data to change (if any).
Your instructor will explain the structure of your csv dataset (Figure 1), which will be different from the Excel sheet you looked at in steps 1-11. In brief, the 24 hr observations of survival are above the 48 hr observations of survival, and each observation in each replicate is situated within its own row; the data are organized in this way so that they can be analyzed properly in R later in this lab. This means that physical information from each replicate, which does not change from one Daphnia observation to the next, is repeated for the 24 hr Daphnia observation, and the 48 hr observation, as exemplified by the arrows and highlights in Figure 1. Note that each replicate has its own Replicate ID, which identifies the replicate within the Excel sheet. The consequence of the data structure has a very important bearing on your lab: if you decide to change any information regarding a physical replicate (e.g., in your Excel sheet, if you erase a conductivity measurement in Replicate ID “Z”, and add in a new conductivity measurement), then you must change the information both rows containing the same Replicate ID, as it represents the same replicate.
Figure 1: Data layout from your lab section’s LC50 experiment for analysis in R.
13. Next, download the file named “R Code for Lab 5.txt” from Quercus. Open up R Studio, and load your LC50 data (the *.csv file) into R, naming your data “dat”. After your data are loaded (and named “dat”), run the code chunk titled “Question 12”. Note that “dat” is just an abbreviation of the word ‘data’, and it is what we named your datafile when generating pre-made R code. You could in theory name your data something other than “dat” when you import it, but then you’d have to make sure you modify the R code we gave you so that it recognizes the alternative name you provided.
- 经济Economics
- 金融Finance
- 会计Accounting
- 工商管理Business Management
- 计算机computer science
- 商业分析business analysis
- 法律Law
- 管理Management
- 市场营销Marketing
- 哲学Philosophy
- 心理学Psychology
- 物理Physical
- 统计学statistics
- 电气工程electric engineering
- 化学chemistry
- 生物biology
- 数学mathematics
- 语言学linguistics
- 土木工程civil engineering
- 工程力学engineering mechanics
- 艺术art
PROFESSIONAL HELP!
在线咨询
立即下单
Wechat：THEend8_ 
Email：