heidi.bio - Amino One / October 2016

In this experiment, we explored the value of consumer-level video technology in understanding and documenting multi-day bioengineering projects. In general, adoption of bioengineering in the home may depend on integrating various types of technology that are found in some homes but not all. Employing video in a bioengineering project is a special case in which the home-based bioengineer may already be familiar with the technology and may already have the technology in use; nonetheless, it would usually require a new camera purchase. Our goal was to determine whether the availability of video data was useful in characterizing the success of an experiment. (Our goal was not to determine whether a specific bioengineering success was possible: that had already been established by many other persons in the past.)

We used the Amino One Desktop Biolab with the "Engineer-it Kit" magenta DNA program. This Amino One was from the earliest production, and had the name "Cricket" on the back. Instead of having a person watch the experiment continuously for two days, we set up a Nest Cam Indoor security camera, aimed at the Amino One culture chamber and the top of the touchscreen, to record the entire outcome of the growth of transformed E. coli. The Nest Aware component offered 1080p video streaming into the nest.com cloud. After the E. coli lifecycle had ended, we made time-lapse videos and also used the Video Clip feature in Nest Aware to extract short video excerpts of important events, such as when the color or translucency of the culture chamber was changing or the pH was changing. The VLC media player and FFmpeg software allowed us to examine these excerpts, graph the touchscreen data, and extract representative still pictures for this report. In addition, we attempted quantitative measurement of the color changes, but this was largely unsuccessful. (We also watched the live video stream during the experiment, because we were usually not physically located at the site of the experiment.)

The two small pictures below show the culture chamber before and after the DNA program had its observable effect. The left picture, obtained a few hours after Grow-it Kit step 3, shows the light-yellow color characteristic of the lysogeny broth (LB) media. The right picture, obtained 39 hours later, shows the maximum observed intensity of the magenta color. Here, the E. coli were expressing the ability to produce a pigment that, in nature, can be produced by coral DNA but cannot be produced by E. coli DNA.

All time values in this report are relative to the beginning of the use of the culture chamber. In other words, the zero hour began when the LB media was inoculated with transformed E. coli and the forward circulation, heat, and air were turned on. For approximately the first 15 hours, there was no significant change in the culture chamber. The bacteria population remained at about 150 billion, and the pH remained at about 6.7. (A later table has more precise data, and explains why the bacteria populations in the text differ from the bacteria populations visible in the pictures.) Also, the initial light-yellow color and clear-liquid appearance were always present, as shown in these two pictures from 3 hours and 14 hours after the beginning:

After 17 hours, the light-yellow color was still present but the culture chamber became more cloudy, although the bacteria population and pH remained approximately the same:

At the 20-hour point, cloudiness had increased. Also, the bacteria population had risen to 300 billion and the pH had risen to 7.0:

The interval between 20 and 23 hours had the most rapid pH change, with a climb from 7.0 to 7.8. This picture at 21 hours shows a very cloudy chamber with 300 billion bacteria and a pH of 7.2:

Similarly, this picture at 22 hours shows a very cloudy chamber with 300 billion bacteria and a pH of 7.5:

At the 23-hour point, the color began to slowly move from yellow toward pink. This was difficult to notice at first. Here, the bacteria population was 400 billion and the pH was 7.8:

After 26 hours, there was again more gradual shift toward the slightly darker pink color. The bacteria population was 350 billion and the pH was 8.4:

At this point, additional LB media was added through the Food Port. The culture chamber was about 45% full before this step, and about 67% full after this step.

Next, at the 29-hour point, it seemed reasonable to label the color as pink. The bacteria population was 350 billion and the pH was 8.1:

A better conclusion about the pink color could be reached after 33 hours. At this time, the bacteria population was 700 billion and the pH was 8.7:

After that point, the bacteria population and pH both rose more slowly. Here, at 35 hours, the population was again 700 billion and the pH was 8.9:

At 39 hours, the bacteria population was 800 billion and the pH was 9.1. Often, a pH value of 9 is considered the maximum for continued E. coli growth. Also, the pink color was near its maximum intensity at this point:

The 42-hour point again had a bacteria population of 800 billion, and the pH had edged up slightly to 9.2. This seemed to be the closest that the culture chamber approached to a true magenta color:

After that point, the pH remained at 9.2 for days, and the bacteria population dropped by about 150 billion per day. We did not do extensive work on data collection in this late stage of the experiment. The pink color faded gradually over time. Additional LB media was added at these time points: 44 hours, 73 hours, and 97 hours. After 116 hours, 15 ml of lemon juice (pictured at the end of this report) was added through the Custom Ingredient port to reduce the pH (Grow-it Kit step 4.2). This apparently increased the bacteria population but did not restore the pink color. The pH remained at 2.6 for days after this.

Here is the main set of bacteria population and pH data. The bacteria population displayed on the Amino One touchscreen often jumped by hundreds of billions over the course of seconds. For the data below (and in the earlier text of this report), the approach was to watch a two-minute video excerpt at the time listed in each table row, and attempt to manually compute a median value of all of the displayed numbers. (The pH values never jumped except during a short time window after initially loading the culture chamber.) Note that LB media was added at the 26.5 hour point - this is correlated with the drop in pH at that time. Our graph was made with GNU Octave:

A table of the same data:

We were also very interested in quantitative measurement of whether the culture chamber had yellow liquid or pink liquid. In attempting to achieve this, we created histograms of the color component distributions in our pictures of the culture chamber. We had data in the YUV color space, and expected that the U component would increase during the transition from yellow to pink. This could potentially be used in an automated process that determined when a sufficient amount of pink was present. Specifically, we would be able to use a Nest API to obtain a snapshot_url value from the "cameras" object, and then visit this URL to download a JPEG image. Next, a histogram could be produced, and further calculations could determine whether enough U values were greater than a specified cutoff value.

In other words, the Nest Cam Indoor product has (at least in theory) the data-acquisition capability for determining the time at which E. coli are expressing the ability to produce a pigment. This could be useful if one wanted to begin pigment extraction at that time. It could also be useful if one wanted to alert interested persons about the pigment, so that they could go to the Amino One's location for a hands-on microbiology learning experience.

We had 1280x720 JPEG images from our camera. We found that a 175x18 image excerpt taken at position 362:56 typically showed a representative picture of the liquid color within the culture chamber. The histograms of U data values from four such 175x18 images are shown below. The upper left is at a time of 3 hours (yellow liquid), and the lower left is at a time of 42 hours (pink liquid). It appears that the U values are generally greater in the latter case, but there is not a pronounced effect. The upper right is at a time of 29 hours and the lower right is at a time of 33 hours: as mentioned earlier, these times were characterized by an initial slight appearance of pink.

We do not feel that this methodology is currently adequate for automated identification of pigment production. One of the challenges is capturing a picture of still liquid without any bubbles. In future work, several JPEG images could be obtained and the histogram data could be averaged with exclusion of outliers. For example, given the overall pace of the experiment, data acquisition with a Nest API could be scheduled for once per hour. Then, a set of 100 JPEG images could be gathered for averaging, with 10 per minute over a 10-minute period. (10 per minute is the maximum snapshot_url rate allowed by a Nest API.) Alternatively, if the Amino Labs team adds programmatic control of the Air and Circulation functions, then Air and Circulation could be temporarily shut off shortly before the Nest API calls.

Time-lapse Videos

These are large (50 Mb) video files that may not work on all Internet connections. Creation of time-lapse videos uses the cloud-based software on the home.nest.com web site, with a very simple user interface. It is not necessary to copy days' worth of video data to one's own computer. Similarly, it is not necessary to have previous experience with a video-editing application.

Here are a few additional notes and pictures from earlier parts of the experiment. More pictures will be added soon.

First, we have a series of comments that apply chronologically to the Amino One instruction manual. We felt that the instruction manual was very good, and are simply listing everything that possibly could have been improved for us or other readers.

• The "General safety rules" discuss eating and drinking near the experiment, but do not discuss the more general issue of smell. For some readers. the smell produced by the experiment is important for choosing a room location.
• In general, we often needed substantially more time per step than specified in the instruction manual (e.g., 50% more time).
• There is a missing word at the end of the first part of the "Setting up your system : continuous culturing" section: it says "one for the tube, one for the" (maybe "air" is the missing word).
• Engineer-it Kit step 1.1 should specify writing on the bottom of the outside of the petri dish. That allows easy reading when the petri dish is in an inverted position.
• Ideally, Engineer-it Kit step 1.3 should address the case where it's not possible to see boiling by looking into the microwave. For example, maybe it is best to stop and open the microwave every 20 seconds.
• Engineer-it Kit step 1.5 should clarify whether it is acceptable to microwave several times, or for a longer period of time, to achieve the non-cloudy state. We needed a total of 25 seconds.
• We did not fully comprehend Engineer-it Kit step 1.10. For example, we did not know whether we should let the LB agar fully harden, or simply observe the suggested "20-25 minutes" time period for all of Engineer-it Kit step 1.
• Engineer-it Kit steps 2.2 and 2.3 seem underdocumented. Previous microbiology experience may be needed to use a loop correctly.
• We considered the Engineer-it Kit step 2.4 "with the lid down" phrase to be ambiguous. Based on step 5.4, we interpreted this to mean "in an inverted position" and did not interpret it to mean solely "with the lid on."
• Also, in Engineer-it Kit step 2.4, it seemed very challenging to place the plate in an inverted position without spilling any of the remaining liquid parts of the agar. This issue probably would not have occurred if the LB agar had hardened for a longer time.
• For the various tubes used in Engineer-it Kit steps 3 and 4, it could be helpful to state what percentage full they are supposed to be, and their viscosity. Readers unfamiliar with microbiology might incorrectly fear that they are empty.
• Engineer-it Kit step 3.3 seems underdocumented. Previous microbiology experience may be needed to understand the concepts of "separated colonies" and "on the tip."
• In Engineer-it Kit step 3.5, it unclear how one would "see clumps" if a tube is in a tube cooler. The reader may erroneously conclude that all of steps 3.4 and 3.5 should be done with the tube outside of a tube cooler.
• The beginning of the Grow-it Kit instructions should inform the reader that the Grow-it Kit should be used only if one is sure that there will be time to complete the entire "Disposal & Clean" process within about a week at most. Keeping circulation on for multiple weeks after step 3.9 is not optimal, because this increases the risk of debris in the culture chamber that may be very difficult to clean. Similarly, step 4.4 should be revised to suggest a reasonable upper bound on "as long as you want."
• An appropriate Grow-it Kit step, such as 1.1 or a new step 2.11, should remind the reader that the Inactivation Tablet must be in the waste container. The "Clean-it Kit components" section mentions "The Inactivation Tablet is Amino Labs third stage of inactivation. The Inactivation Tablet is placed in the waste container just prior to culturing and is active for 5 days." This reminder is too late. There is an earlier instruction under "Setting up your system : continuous culturing" of "drop an inactivation tablet from the Clean-it Kit inside the waste container" but this does not state when that should happen, i.e., it omits the "just prior to culturing" information.
• For the Grow-it Kit, it would be helpful to identify steps that can be performed concurrently. For example, while one person is completing the Short Clean steps, another person can be doing steps 2.1 through 2.4. This can save a lot of time if 25 minutes are needed for the pill to dissolve.
• We found that it was sometimes difficult to understand or complete connections to the ports, e.g., in Grow-it Kit step 1.10.
• In Grow-it Kit steps 1.11, 1.12, and 3.6, we experienced a minor amount of spillage because the port size and syringe size did not match closely enough during injection. This was also an issue for the Clean-it Kit (e,g., steps 1.11, 1.12, 1.21, 1.22, 1.30, and 1.31).
• In Grow-it Kit steps 1.19 and 1.21, we were not able to draw water from the rinse bag; instead, we drew from a gallon container of distilled water.
• A few steps such as Grow-it Kit step 1.22 and Clean-it Kit step 1.1 refer to a "dispose" button; however, "Discard" is the correct name.
• We did not achieve the value of 4 in Grow-it Kit step 1.28. We stopped that cleaning stage after reaching 3.2.
• It would be helpful to state a motivation for Clean-it Kit step 1.18. In other words, measuring exactly 30 seconds will not be optimal for every reader; the desired outcome is to have a small amount of water in the culture chamber, e,g., maybe about 10% full. Possibly the goal is to clean out debris from the bottom surface of the culture chamber.
• Clean-it Kit step 1.23 refers to 3 minutes but this seems inconsistent with step 1.19.
• Various Clean-it Kit steps refer to steps numbered with a "5"; however, all steps start with a "1" instead, For example, Clean-it Kit step 1.27 mentions step 5.20, but is actually referring to step 1.20.
• Two different Clean-it Kit steps are numbered 1.32.
• The antibiotic was labeled "kan" - we expect that it meant kanamycin.
• Other than the noted connection problems and instability of the bacteria-population values on the touchscreen, there were no known operational problems with the Amino One. For example, it did not ever need to be turned off and restarted.
• The Amino One WebApp was, as far as we know, not available at the time of this experiment.

Next, we have pictures of some of the significant parts of the experiment, also organized chronologically.

Engineer-it Kit step 2.4 - growth of non-selective plate after 6 hours:

Engineer-it Kit step 2.4 - growth of non-selective plate after 14 hours:

The non-selective plate after Engineer-it Kit step 3.3:

Engineer-it Kit step 4.3 - incubating for 20 minutes at 4 degrees (or slightly lower):

Engineer-it Kit step 4.4 - achieving 42 degrees:

We used the time.gov web site to measure the critical 90-second period for the heat shock in Engineer-it Kit step 4.5. Beginning of the 90 seconds:

During the 90 seconds:

After the 90 seconds:

After Engineer-it Kit step 4.6 (3 minutes at 4 degrees):

Engineer-it Kit step 5.3 - selective plate:

Engineer-it Kit step 5.4 - incubation of selective plate at 37 degrees:

We aimed the Nest camera at the incubator so that we could record the growth and watch it remotely (we found that the zoom and enhance features in the Nest app were helpful):

Engineer-it Kit step 5.4 - growth of selective plate after 6 hours:

Engineer-it Kit step 5.4 - growth of selective plate after 14 hours:

Engineer-it Kit step 5.4 - growth of selective plate after 26 hours (the plate had been moved slightly when looking for colonies):

Starting after Grow-it Kit step 3.9, we moved the Nest camera to a higher position so that it could see both the culture chamber and the top of the touchscreen. We simply stacked miscellaneous household objects to achieve the correct height:

Here is an overview of the complete lab setup including the waste container:

Conclusions

We found that the Amino One Desktop Biolab (aka Bioproduction Lab) is a reliable platform for DNA transformation experiments. We believe that capturing a complete multi-day video stream of an experiment is extremely helpful for ensuring that many types of data and observations can be reliably collected on an after-the-fact basis. We do not have any personal interest in specific camera manufacturers, but we feel that the Nest Cam Indoor product is a robust and cost-effective choice. This is marketed for home consumer use, and is thus compatible with the home-based bioengineering concept. The entire process of integrating video technology into a biology experiment has non-negligible cost and complexity, but adds substantial value to the educational outcome.

Extra Pictures

Here are some of the additional products we bought for the experiment. First, we used a store brand of lemon juice to reduce the pH:

We found that the one-gallon Rubbermaid mixermate pitcher was a good choice for the waste container:

We found that a small folding table could be used to place this container's opening just below our countertop: