"
Team:Wellesley HCI/Notebook/NicoleNotebook
From 2012.igem.org
| Line 128: | Line 128: | ||
</script> | </script> | ||
| - | + | </html> | |
<br> | <br> | ||
| Line 248: | Line 248: | ||
<!--notebook ends here--> | <!--notebook ends here--> | ||
| - | |||
Revision as of 05:41, 10 July 2012
Nicole's Notebook
Day 1: May 29, 2012
Research started today, so a majority of the day was dedicated to administrative details about the program. Otherwise, Veronica and I were assigned to research more about semantic search--what it was, its benefits, and whether or not it would be worthwhile to pursue this field in the context of synthetic biology. We will be presenting our findings near the beginning of June.
Day 2: May 30, 2012
For the majority of the day, I was learning how to code with C# on a Windows 8 Surface. It was easy to get lost on the more technical side and details of the coding, especially because the instructions seemed to assume you knew about the basic commands in C#. Other than that, it was easier to delve into the code without blindly copying down if it had been my first time. There were a couple of times where getting stuck and being unable to figure out the problem because of my lack of knowledge got frustrating, but because it was my first experience creating code to manipulate objects on a touch screen, I found the learning experience to be pretty amazing overall. Otherwise, I’ve been doing more reading on synthetic biology. It took me a while to read them since I hadn’t taken up Biology since the ninth grade, but I found the idea of engineering a new biological function interesting so I managed to complete most of the reading.
Day 3: May 31, 2012
I continued reading more about Semantic Search and its advantages compared to other types of searches as well as its disadvantages along with exploring different semantic search engines that are already available on the web, specifically for bioinformatic research (e.g. [http://bio2rdf.org], [http://distilbio.com], and [http://nlmplus.com]). A couple of papers also introduced me to the idea of incorporating other types of searches in Semantic Search. I learned a bit more about text mining, natural language processing (NLP), and ontology.
Day 4: June 1, 2012
Day 5: June 4, 2012
Veronica and I continued reasearching on Semantic Search for the morning. The lab then headed over to Microsoft NERD for the second half of the day where we learned about their new beta project, a tablet with Microsoft 8. The interface is definitely engaging, what with their idea of using tiles and semantic zooming. I also like how transparent they’re trying to be with people who are trying to create applications for their story. Their idea of using a picture the owner can choose on their own as part of the password was really cool! I also like how they found a way to accommodate different finger sizes by enabling users to rescale the keyboard. All the different gestures to fully use the tablet however are not intuitive and takes some getting used to. It seemed much easier when Edwin was just demonstrating the product for us than when I actually tested out the tablet myself. Although, it is great how applications can now be made using HTML5 and JavaScript. I'm definitely going to start learning how to code in these languages before the summer is out.
Day 6: June 5, 2012
The lab met up in MIT and we spent the entire day learning more about synthetic biology. Professor Natalie Kuldell gave us a lecture in the morning and had us running experiments that afternoon to try applying what we've learned and get us exposed to the lab environment (which was particularly helpful for those of us who hadn't done Biology since 9th grade). Natalie was able to make the lecture engaging when talking about the basics of synthetic biology and its similarities with other engineering practices. Once she got more technical however it was more difficult for me to follow. As for the lab, first of all, I've definitely gained a newfound respect for people majoring in the biological sciences. Lab can be very tiring--and we were only in there for two hours when labs in Wellesley College courses usually take 4+ hours!
It was difficult to remember reasons behind you're doing the protocols you're doing especially if, like me, one is a beginner in the subject who's more immediately concerned with what the lab procedures mean by using a certain instrument or machine to complete a protocol. It was helpful that Natalie had led us into a discussions after the lab to reflect on what we just did and sort of touch upon the big picture, but I definitely felt like, by the end of the day, I learned more about lab processes rather than the material that motivated the experiments in the first place. Natalie highlighted how difficult it is to look for information when the researcher does not know what to look for--something that hopefully we'll be able to facilitate using the applications we'll be developing for iGEM and semantic search.
Day 7: June 6, 2012 - Day 9: June 8, 2012
Veronica, Casey, and I discussed our individual findings from the research we had done until this point and decided what we'd be presenting -- the basics of semantic search, ways we can implement semantic search in surface applications, and challenges of implementing semantic search. We also tried finding an already made API we could customize because creating our own semantic search engine is definitely not a feasible job for a few people in one summer.
We spoke with Consuelo and Orit about our current material for our presentation this coming Monday and found that we needed to research more on the feasibility of implementing a customized semantic search engine using databases like PubMed, iGEM, and the Parts Registry. We then finalized our presentation.
On Friday, we were able to speak with Professor Eni Mustafaraj. She went over a brief explanation of semantic search and how it is implemented. She gave us suggestions on ways we could have our own customized semantic search engine. Initially, we had thought of using the commercially-made engines that enabled users to modify it based on the databases they want the engine to use. Eni warned us however that we should search for semantic search engines that had been written about in a research/academic paper, and instead pointed us towards an open source project called Apache Jena. What we would have to do is somehow convert what we want into RDF format and from there figure out how we could implement the engine using Jena.
Day 10: June 11, 2012
Today, each group presented their findings on the topics we had been assigned at the beginning of summer research. Between each presentation, we shared ideas that popped into our heads and devoted around 10 minutes to brainstorm how each of what had been researched could be integrated into our iGEM submission.
I found the idea of the DiamondSpin (i.e. the Lazy Suzanne for platforms) really interesting. It has several functions but I found it to be most intuitive for the windows to shrink as they were dragged towards the center. Then, should a person at the other end of the Beast, for example, want to access it, they would just use the outer rim of the circle to rotate it such that the window’s orientation would be adjusted appropriately and it was closer to them. Then, they’d be able to drag it outside the circle and examine it. As for the idea of windows getting larger as they were dragged towards the center, while this may seem useful for presentations, considering that the circular form makes sense for a platform with people surrounding it like the Beast (change in orientations), then enlarging it in the middle to present it to people would keep it in one orientation. Consequently, the orientation may be readjusted enough times that it would become a distraction during the presentation.
After the presentations, we were split up into two different groups and each came up with five different big ideas for projects to pursue over the summer. I really liked the idea of having a graph, inspired by the idea of Semantic Search, that essentially tracked the person’s “trail of thought.” They begin with a main query that constitutes the middle bubble. Then, the succeeding nodes represent queries they made right after and/or related to the node it is attached to. If the person wanted to go back to change a query or start another trail, they’d be able to. There was also the idea of the personal lab helper (like the animated clip on Microsoft Word). The program would different between people by recognizing their irises or fingerprints to prevent confusion. The surface would be the lab table and through each protocol, if a certain tool was needed, the shape of the tool would appear on the surface and keep blinking until the person prepared the correct tool. A camera would be placed above the table and record while the researcher would be doing an experiment. There were also a lot of ideas for the eLab Notebook, like being able to zoom in to a protocol and annotate it but when you zoom out again, all that has changed is that the protocol is now either highlight or changed its color to red. The user would then know that they had annotations and there would be no clutter.
Day 11: June 12, 2012
The Wellesley Lab went over to Boston University. We were able to meet the two undergraduates who are part of the BU team. Traci presented more about Synthetic Biology and went into more details about parts names, what they did, and we learned more about what they are currently doing for the competition. Afterwards, three from the Wellesley team presented our research topics and the ideas that resulted from our brainstorming the day before. We then proceeded to brainstorm even more, gaining valuable insight from the BU team.
Day 12: June 13, 2012 - Day 13: June 14, 2012
Main brainstorming started today. In the morning, we set up a classroom such that each project had a different section. We stuck photos associated with our projects on the walls, and we wrote on post it notes and on the whiteboard of the ideas we currently had. When the BU team came over for the day, we went around to the different sections and had a huge group discussion for each project with the original subgroups leading the discussions. We talked about the possibilities of each project, any ideas pertaining to the project that popped out at other people, examples of similar projects done in the past, other things (e.g. platform applications) that can be incorporated into them (e.g. for the beast: DiamondSpin). We learned more about what the art project entailed and clarified the type of audience it would be aimed towards. Natalie Kuldell was able to stop by and gave us very useful input on what she thought of our current ideas for the eLab Notebook. Two things were made very clear: being able to search through the notebooks a scientist has accumulated throughout the years of their research would be a valuable tool; and, having the freedom to write whatever you want on the notebook is important, so we’ve started researching on styluses we could use.
Day 14: June 15, 2012
More research on pens, dump registry to XML and convert to RDF format for Apache Jena, transcribed audio files, terms and conditions of different databases like GoWeb, Google Scholar, etc.
Day 15: June 18, 2012
Project status update. Currently finding it difficult to dump the database into XML and converting it into RDF because of the huge amount of data—might try to separate the site into smaller parts. While we are not very hopeful with our group’s part of the research because of the size of the database we’re trying to crawl through, we still haven’t exhausted all options and so are continuing with the research. Trying to look for anything similar to what we’re trying to do that had been done in the past, either by previous iGEM teams or not. Reading through the team abstracts from 2008, we weren’t able to find any team who had tried to convert the Parts Registry into RDF format or, in general, tried to create a search engine for it, semantic or otherwise. Veronica did however manage to find something called Knowledgebase of Standard Biological Parts, “a Semantic Web resource for synthetic biology” which managed to transform information from the Parts Registry to make it computable. This seems to be what we are essentially aiming to create, so investigating more on how the this resource was created will hopefully help us progress from our current rut. If we managed to create a structure for the Parts Registry, we would be able to access the information from individual teams’ notebooks. Trying to mine just part of the data (breaking it up into smaller collections), building own crawler, what else is necessary to customize our own engine.
Day 16: June 19, 2012
Planned out the rest of the summer → starting to code/do technical feasibility checks. Initially, started research on web crawlers that pulled data from the Parts Registry website for the idea of having a semantic search engine for the Parts Registry among other things. However, because the process of trying to use Apache Jena is difficult, Conseulo suggested to use a Google Custom Search Engine through C# and have results pop up on a window. Managed to implement Google Custom Search Engine using C#, which is displayed in a DataGrid. Casey managed to code up a server which enables him to drag an image file from a tablet to a window and have it appear on the desktop computer. Linda managed to figure out a format for a data sheet for a part. Going to work on testing out how to search through PubMed and iGEM. Will check the quality of the custom search results compared to Parts Registry ‘Text Search’ results.
June 20: (Day 17)
Comparing Google Custom Search results between what shows up in the C# implementation and when testing out the custom search engine in the site, there is a significant difference. For one thing there are only 10 results that appear in the pop-up window of the C# code whereas there are 10 pages worth of results in the sample site. Moreover, the results do not show up in the same order, so now we are trying to figure out how to make the results that appear on the DataGrid window be more similar to how they should appear. To be able to search through PubMed, Consuelo provided us with the web crawler she coded for one of last summer’s projects. It definitely works with genes, so we’ll have to test out whether it works for phrases or other keywords. As for searching through the iGEM archive, we’ve added all previous iGEM sites to what the custom search engine will look through.
Unfortunately, we weren’t able to figure out how our problem with C# implement of Google custom search results in terms of order and amount, so we’ve decided to customize Conseulo’s previous web crawler for PubMed and modify it for the Parts Registry. So far, we’ve managed to find a URL scheme for text searches. We are learning the code behind her Web Crawler before we start editing it to better suit our needs. With the web crawler, we’ll also be handling the results that appear by going through the links and creating data sheets from the links to the different parts.
Day 18: June 21, 2012
Veronica and I managed to finish customizing part of Conseulo’s web crawler, so now we have a Web Crawler that goes through the search results of a query and extract the title of the pages, the links to said pages, and the number of results (which appears if the query is not too generic). Linda sent us a document listing the information she needs to be able to construct a data sheet, and we’ve come up with a structure for the classes needed to be able to extract the information in an organized manner and are currently working on the code.
Day 19: June 22, 2012
Consuelo sent us another database from which to extract information to create a more complete parts datasheet: BioFab.com. BioFab’s projects are aimed to sharing information on the parts they themselves develop, but looking through how they organize this data gives us another idea of how our datasheets can look. Unfortunately, their database, C-dog is not searchable. They instead list out the and instead organize them by their name, construct, description, mean fluorescence per cell, and standard deviation. Choosing to organize them by name though is a bit frustrating. Each of their parts has the same prefix, “apFAB” followed by some number. Unfortunately, because they don’t precede these numbers with 0’s to ensure that 47 is not considered greater than 101 because the leading digit, 4, is greater than 1, then the order is not correct. However, I assume they provide information on standard deviation and mean fluorescence per cell because they are a major factor in helping biologists decide which part they need. We’ve started to finish up creating a more comprehensive data sheet using the information from the Parts Registry as listed by Linda in a document. However, because the pages are incomplete we are unable to extract some of the points. We still have to work on Web Crawlers for other databases, like BioFab and others, and customize the old Web Crawler for PubMed to extract information on parts instead of genes.
Day 20: June 25, 2012
Until we get a response from our source about where to obtain more of the information we need to complete the data sheet, the code we have so far is essentially complete. Veronica and I managed to finish debugging and cleaning up our code (commenting, creating properties, etc.). This version of the crawler is only able to handle Parts pages. Because we were able to finish debugging before the end of the day, we try to use the HTML Agility Pack to try to parse through the other types of pages and try to extract all the content within the HTML and CSS style tags. Unfortunately, we ended up in a dead end because the resulting output was a mess due to the page’s formatting. Some pages just had headers and text, which was easy enough to interpret and if we had spent a bit more time we could have tried making the output relatively more readable for the reader. But, a lot of pages contained images, captions to these images, tables, and other representations of information, so the subsequent output was too much of a mess to read through. Instead, we’ll just have the Crawler process Parts pages, so next time we have to work on making sure that the rest of the code does just that (e.g. editing some classes to ensure that the lists containing the results to the search queries only contain links to Parts pages).
Day 21: June 26, 2012
Started to connect RegList.cs and RegDataSheet.cs via a method called search that takes a string (the query) as an argument. The first class parses through the search results page whereas the second crawls through the actual results pages. We need to connect it such that the Crawler invokes the RegDataSheet.cs for each link. A lot of troubleshooting needed to be done here, and difficulty was mostly caused because we hadn’t taken all cases into consideration. We’ve continued debugging by looking through the output of our classes and the actual information presented in the pages.
Day 22: June 27, 2012
Encountered problems with algorithm that parses through the search results page, so we spent the morning trying to solve it. Worked with Linda on connecting our back end of the MoClo Planner to the front end she is developing using HTML5 and JavaScript
Day 23: June 28, 2012
Reformatted the string returned by RegDataSheet in the order Linda wanted. Continue to find ways to extract other necessary information from the Parts Registry.
Day 24: June 29, 2012
Had a bit more difficulty with extracting the assembly compatibility but was able to complete the working code in the morning. Worked on making sure Parts Registry Crawler was commented and understandable. Started working on modifying Pub Med Crawler for our current purposes.
Day 25: July 2, 2012
Integrates PubMed Crawler -- instead of being able to find information on specific parts (because no results show up when one does that), just use a general text search and extract titles, authors, and abstracts. Worked on fixing the combination of the Results Page and Parts page parsers.
Day 26: July 3, 2012
Our summer interns from Framingham High School came in today! Two rising juniors, Maura and Madeleine. Linda and I started teaching them HTML and they started working on making their own website. In the morning, Casey and Consuelo held a tiny boot camp to make sure we were on the same page with coding and the Windows Application. Finished debugging crawlers and ensured that the output .txt file that the C# program creates is in the format Linda is able to work with. Created an executable file. Had to transfer it onto a new Console Application project so that Linda can access the code using a command line argument and through it populate her online data sheets. Have to read up on cgi-bins and how to create them for my code. Better commenting the code for Nahum and Kara to be able to use through getters.
Day 27: July 5, 2012
Was able to finish creating an executable/console application for the PartsRegBrowse code. Will continue work on coding up a CGI once Linda finishes the user interface for the front end. Debugged a bit more—modified for cases that had previously not been taken into account. Switched onto the Sifteo project so I talking with Kim about what she had currently coded up for the cubes and started learning about Sifteo API. We were able to make progress with a problem in the program. The colored cube can now drip into a neighboring plasmid cube! The next step is to be able to do that with whatever color the user decides on (i.e. red, yellow, or blue). Unfortunately, as we were trying to solve this next step to the problem and decided to restart from our last big progress (i.e. the dripping), we had found that that last update had not been saved onto the repository, so altering the code took longer than expected.
Day 28: July 6, 2012
Was able to get the animation to work with the other colors (blue and yellow). Met with Orit today to figure out the next step for Sifteo cubes so that we can have a demonstration by the end of the upcoming week. Have to include the E. Coli cube and code up how it appears alone and how it interacts with the others
Day 29: July 9, 2012
Started coding up the third (E. Coli) cube and trying to figure out how to continuously run the animation with the least amount, ideally without any, prompts. Currently having trouble with using the actual pictures for the dripping animation. There is more than twice the amount of frames for the finalized animation than there was for the animation created to test out the code. Not sure if it's because of the code (which should work regardless of the number of frames) or if it's because of the SiftDev/Siftulator App that images are not appearing on the simulated cubes. When Kim tested the finalized animation on the actual cubes, they worked with an error coming up (which we still have to figure out). Hopefully we'll be able to figure this out by tomorrow because it'd be great if I could concentrate on getting the E. Coli cube to work (including the main functionality behind it which is have it interact with the colored cube and, if said cube is flipped, color will transfer onto the E. Coli cube).
Really hoping to get all this done by Friday since Demo session with high school students is on Wednesday. Devoting Monday and Tuesday to debugging any more (hopefully minor) problems with the demo.
