21/7/2015
Today is our first day after the conclusion of the bayram, so obviously our crew was a little sluggish to begin the day. Once we were up to speed we began the day differently with a short building session in our Minecraft server involving the stadium again. The additions consisted of adding nice looking pillars to take the weight of the seating area, refined fences which allow a walkspace near them, and flags for the corners of the football field.
We moved on quickly to write some code relating to the CFTR sequence we were working on last week. This time we had to take a motif and give the end result with all the number of ‘N’s for each motif added together to give the most common motifs regardless of ‘N’ size.
We quickly moved on to hear a bit about chapter 8 of Data Science from Scratch, which involved gradient descent and how we should choose the right step size. Also it mentioned the stochastic gradient descent as well.
To start our night lesson we had a face to face presentation with Onur Kaya, who is a PhD student at the Institute of Molecular Biotechnology at Vienna. He is working with Drosophila to further understand cancer cell replication. He mentioned that getting any sort of experience from out of country is a vital part of any CV and employers tend to focus on it. Tying into this he told us about his internship in Germany and how it helped him to get his current employment. Then he said that knowledge in biology is necessary in his line of work and they require bioinformatics to analyze the mass amounts of sequences that they produce. He mentioned that if we talk about problems through a mathematical window instead of a biological window we start to see gaps in the data, and that's why people with biological knowledge was key. Moving on he said how he wants to save time by using programming languages like R to achieve his goals. Also the maximum working hours a week in Vienna are 50 hours but overtime can rarely happen, but people he works with all try their best to solve the problems they face. To repeat a previous point he told us that we should have a full CV and must know how to show what we know to potential employers as he knew people with great potential that were unable to transfer their knowledge into something an employer would take interest in. Then he remarked on how most institutes need a big IT team due to the huge networks they have, so computer engineers are important in molecular biotechnology. Looking back on the topic of the internship he got in Germany, which was Chromatin Immunoprecipitation assays (ChIP assays), he finds that ChIP assays take up a majority of his activities in his new job and reiterated the importance of doing an out of country internship.
To end the day we talked about what we have been learning so far in our internship and everybody agreed that they had improved themselves in one way or another. We happily ended the night with laying out our plan for the coming days/weeks by listing the tools we would be using which included D3 and Django.
22/7/2015
The day was young as we entered the building early for an interesting team building exercise where we played the popular game called agar.io with the goal to collect nutrients for the main cell in our team of cells to get the highest possible score. Despite our best efforts we failed to reach first place but came close many times.
Following our morning adventure Mr. Ahmet gave us a general overview of our plan for the next three weeks. He started with an overview of recently occurring bioinformatics seminars including HIBIT2015 and MEMBS and told us it would be nice to attend them if we had the time. He then went into detail about protein enzymatic functions and their interactions with substrates. Before going into more details he gave an example from the previous years bioinformatics intern group where they took writing samples from everyone in the group and compared them to an unknown sample of text to determine who the writer could be. This was possible by using Bayesian probability. He recommended a book called Think Bayes, which is the book he used to learn Bayes’ theorem. To tie in Bayes he talked about protein docking, which is when proteins interact with other proteins to understand how their 3D structure would look like in living organisms. Small proteins can bind to many other molecules other than their intended binding protein, and this can give rise to side effects in drugs if our drug binds to an unintended protein. We use Bayesian probability to compare proteins that we don’t know the functions of to proteins we do know the functions of to ultimately learn more about how they might interact with other molecules in living organisms. Even then we still don’t fully understand all the 3D structures that can occur. We can also use ChIP sequencing to map out protein binding sites with DNA. There was also mention of using FASTQ which has high reading quality to avoid reading errors for the stored sequence.
Next up we moved on to learn the mechanics of another teambuilding game - Game Dev Tycoon. We would have to get together in groups and try to make the most money in a set amount of time. The challenge would be that the contribution of each team member could adversely affect the outcome of the game.
Nearing the middle of the day we were left with internet access problems as there was a massive internet upgrade session going on to implement further security. Since our internet connection would be gone for rest of the day we had to resort to the non-internet based duties of the day.
We wrapped up with some Python knowledge coming from Chapter 9 of Data Science from Scratch, which was all about getting data. The chapter consisted of explaining piping data, reading text files in Python, delimiting files, scraping the HTML data of websites using the BeautifulSoup library, and using APIs via JSON and XML.
.......
This concludes the beginning of our shortened week, stay tuned for more!
Hiç yorum yok:
Yorum Gönder