An introduction to the statistics behind the most popular genomic data science projects. This is the sixth course in the Genomic Big Data Science Specialization from Johns Hopkins University.
Getting Help (3:44)
Loading...
来自 Johns Hopkins University 的课程
Statistics for Genomic Data Science
129 个评分
从本节课中
Module 1
This course is structured to hit the key conceptual ideas of normalization, exploratory analysis, linear modeling, testing, and multiple testing that arise over and over in genomic studies.
与讲师见面
Jeff Leek, PhDAssociate Professor, Biostatistics
Bloomberg School of Public Health
Statistics for Genomics can be a pretty complicated topic and so
you're probably going to need some help at some point.
This lecture's a little bit about what are the fastest ways and
the easiest ways to get help.
The best way is going to be the course forum.
The course forum is a place where I can see all your questions that you're posting
and you can also get help from community TAs who are people who've
previously taken the class and are taking it again to try to help you out.
You can also get help from your fellow classmates who are taking the class
at the same time.
Outside of the course,
there are a couple of other good places to go to ask questions.
One is you could go to the support pages for Bioconductor.
A large number of the things that we'll be talking about in this class
relate to Bioconductor R packages.
And this is a great place to post your questions and
get feedback relatively quickly.
Another good place to go is to Stack Overflow.
Stack Overflow is another place that has general purpose questions that you can ask
and if you want to ask questions about our Bioconductor you can add tags
that will label it question and hopefully you'll get answer relatively quickly.
If you want to get an answer very, very slowly you can send me an email.
I really discourage people from sending me direct emails.
It's probably mostly because if I get all of the questions in my inbox,
it'll take me a huge amount of time to answer them.
And since I am only one person, and
it tends to be a lot of people taking the class,
at any given time, it'll be really slow for me to respond.
I do try to respond to those questions as often as I possibly can, but
it will be probably the slowest way, unfortunately, for you to get an answer
because other people won't be looking at it and giving you answers more quickly.
And I know I like to get the answer almost immediately when I'm posting questions.
A couple of things that might be helpful when you're asking questions.
So if you're asking R questions, there're a couple of things you might think about.
First, you want to report what are the steps that you did to get to that problem?
What are the software packages you used?
What are the functions that you used?
Explain a little bit about what you were trying to get and what you got instead.
That way people will be able to go through and try to reproduce what you did and
see if they can figure out the problem.
You should also try to post what version of the product you're using.
For example, dev tools gives you the session information.
And you can use that in order to report what are all the packages you're using?
What are all their versions?
What's the version of R and what operating system you're on?
Are you on Mac or Windows?
Unfortunately some of the things are a little bit different between the platforms
and so it's helpful to know what platform you're working on.
Similarly for data analysis questions it's good to
report what's the exact question you're trying to answer.
What are the steps and tools you use to answer it?
And then it's really important to report what you expected to see so
people will know what the output, why it doesn't match what you expected.
And then, what did you see instead?
And it often helps to report a few of the other solutions you've thought about when
asking the question.
That's just so that
people don't immediately jump to an answer you've already tried and it didn't work.
Something to keep in mind is whenever you're doing an email or
post it's really important that you stay friendly and
positive, especially in the forums in a big class like this.
You want to make sure that everybody has a good experience in the class so
let's try to stay friendly and positive and constructive on the forums.
Some bad posts or email characteristics, are that they are very vague.
They don't explain what the versions are.
They don't expect what the questions are.
So, help, I can't fit a linear model isn't very helpful.
A better question will give you the version number of say the software that's
being used, and what the specific problem is like that there's a seg fault with
a large data frame.
And it's even better if you can be more concise about it.
So, if you can tell exactly what the function is, exactly what the software is,
the sparsest amount of information possible, that's the best.
So, it's kind of trying to balance giving as much detail as you can
with balancing that out with making sure that you don't write a huge
pages long post that people can't get through.
So, hopefully that'll help you get answers to your questions as fast as possible.
