search

hackseq / vasco

Visualization of single cell RNA-seq data from tens of thousands of cells in real time
6 stars 3 forks source link

Welcome to the single cell visualization team! #1

Closed gracezheng closed 7 years ago

gracezheng commented 7 years ago

Hi everyone,

My name is Grace, and I work on single cell applications at 10x Genomics, a biotech startup in the Bay Area, California. I lived in Vancouver for 5 years, and went to UBC for my undergrad. I am really excited to work with you on the single cell visualization project in Vancouver this Oct.

It'd be great if you can introduce yourself and your background, so that I can provide some reading material for you. In the meanwhile, please take a look at this site, http://support.10xgenomics.com/single-cell/software/overview/welcome, and this tutorial, http://s3-us-west-2.amazonaws.com/10x.files/code/cellrangerrkit-PBMC-vignette-knitr.pdf. While we are not going to write up a R package for this visualization project, I think the tutorial illustrates some of the use cases that we may want to consider. In addition, I really like this website for its way of presenting some features of the neuronal single cell data, http://casestudies.brain-map.org/celltax. Maybe we can get some inspiration for our visualization project. Please let me know what comments, suggestions, and questions you have!

The data I am thinking of using is a 2700 PBMC dataset we have collected at 10x Genomics, http://support.10xgenomics.com/single-cell/datasets/pbmc3k

Thanks, and I look forward to starting the discussion now, and meeting you all in Oct.

Grace

1234jc4321 commented 7 years ago

Hi Grace and teammates,

I’m really excited to be part of this visualization project.

My background is mostly in genomics of respiratory diseases. I’m used to developing scripts to analyze genetic variations.

I work with R to generate most of my figures.

I never had the chance to work on single-cell genomics before. I’m looking forward to helping the team and learn from this challenge.

Have nice evening,

JC

On Sep 14, 2016, at 9:21 AM, Grace Zheng notifications@github.com wrote:

Hi everyone,

My name is Grace, and I work on single cell applications at 10x Genomics, a biotech startup in the Bay Area, California. I lived in Vancouver for 5 years, and went to UBC for my undergrad. I am really excited to work with you on the single cell visualization project in Vancouver this Oct.

It'd be great if you can introduce yourself and your background, so that I can provide some reading material for you. In the meanwhile, please take a look at this site, http://support.10xgenomics.com/single-cell/software/overview/welcome http://support.10xgenomics.com/single-cell/software/overview/welcome, and this tutorial, http://s3-us-west-2.amazonaws.com/10x.files/code/cellrangerrkit-PBMC-vignette-knitr.pdf http://s3-us-west-2.amazonaws.com/10x.files/code/cellrangerrkit-PBMC-vignette-knitr.pdf. While we are not going to write up a R package for this visualization project, I think the tutorial illustrates some of the use cases that we may want to consider. In addition, I really like this website for its way of presenting some features of the neuronal single cell data, http://casestudies.brain-map.org/celltax http://casestudies.brain-map.org/celltax. Maybe we can get some inspiration for our visualization project. Please let me know what comments, suggestions, and questions you have!

The data I am thinking of using is a 2700 PBMC dataset we have collected at 10x Genomics, http://support.10xgenomics.com/single-cell/datasets/pbmc3k http://support.10xgenomics.com/single-cell/datasets/pbmc3k Thanks, and I look forward to starting the discussion now, and meeting you all in Oct.

Grace

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/hackseq/2016_project_7/issues/1, or mute the thread https://github.com/notifications/unsubscribe-auth/AHt470nSlTbFqk8-kA6JJm1nnfibudsiks5qp_ThgaJpZM4J8w7q.

oganm commented 7 years ago

Hi everyone,

I am a PhD student at a neuroinformatics lab mostly working on microarray expression data and some single cell data (including the one in the link above). My entire analysis pipeline is based on R and if we stick with R and want to some web development I have experience in shiny as well. If we don't stick with R I last touched python about 3 years ago, matlab 4 years ago and java 7 years ago but I can try to pick them up if needed.

esmason commented 7 years ago

Hi guys,

My name's Ed and I'm currently a medical student at UBC. I don't have a background in bioinformatics, but I do have a background in both CS/programming and in biological sciences research from my undergrad. I'm very excited to learn more about bioinformatics during the hackathon!

My primary language is python, but I am also familiar with java, MATLAB, c++ and JS (not so much R unfortunately but if we decide to use it I could definitely do some tutorials before the hackathon! ). I have some experience in web development (django, jquery html css) as well as some of the very basic data visualization tools in python (matplotlib).

Looking forward to meeting and working with you all,

Ed

csiu commented 7 years ago

Hi everyone,

I'm a MSc student working on a thyroid epigenetics project using ChIP-seq and RNA-seq data. I have experience with R (+shiny) and python. I am not familiar with MATLAB, but am interested in learning c++ and JS (on my to do list... but haven't yet found time to do so...). Visualization is cool and I also look forward to meeting everyone and tackling this project.

esmason commented 7 years ago

Nice to meet you all,

@gracezheng I just watched a bokeh tutorial and checked out some of the docs. It looks really cool!

I was wondering if we are still planning on using this library because if so, I will start playing around and learning more.

-Ed

gracezheng commented 7 years ago

@esmason, I'm glad that you're excited about bokeh.

I'm actually still debating between bokeh (https://demo.bokehplots.com/apps/stocks, python) and plotly (https://plot.ly/r/shiny-coupled-events/, with javascript or R Shiny). I think both should be able to cover the most important features for this single cell visualization hackathon. I'm trying to evaluate how easy it is to get something running in 2 days. Do you, or anyone have any preference?

Here's the list of features I'm thinking about,

Most important

Nice to have

oganm commented 7 years ago

Personally I'd prefer R as time might be a little tight to re-introduce myself to python though I do see that we have at least one person that would say the same thing about python. Most stuff you say is rather easy to do through shiny though. Below is a link to something I am working on for my project that has some of those already. http://neuroexpresso.org/

esmason commented 7 years ago

I'll have to do some more reading and look at the docs in more detail but from your links and briefly looking at the github, stack overflow etc it seems like.

Plot.ly:

Bokeh:

@oganm I was going to bring this up too because it seems like overall we have more experience with R in this group. I'd love to use python but I've also heard R is not too hard to pick up and I suspect there may be some javascript/html tweaking to do in the front end anyway. Also note both of the proposed libraries have R apis.

@gracezheng are there any features that Shiny would be lacking? It seems like there is no web socket/streaming functionality in shiny but there is in bokeh and plotly. I'm not sure if this is what you were referring to re: "real-time" single cell visualization?

sorry for the long post!

-Ed

esmason commented 7 years ago

Hi @gracezheng

Is there anything you'd like us to have installed and/or familiar with before Saturday? I know we (may) still be unsure about what framework we are using but are there any auxiliary packages or tools we should have set up?

Thanks,

Ed

gracezheng commented 7 years ago

Hi everyone,

Sorry about the silence. We are getting ready for an exciting ASHG at 10x!

I have more clarity on the framework. Given people's background and interest, I think it'd be the easiest if we go with plotly and R (with Shiny).

You can get a quick tutorial here, https://plot.ly/r/. In preparation for the hackathon on Sat, it'd be great if you can download R (I like Rstudio, but that's a personal preference, https://www.rstudio.com/products/rstudio/download3/), and get the plotly package.

I'm uploading 2 sets of files to amazon (gene-cell count matrix and tSNE coordinates) some time today, and will let you know when they are ready for download. Feel free to play with them. I can show you how to read them into R on Sat. If you need to get some background on single cell RNA-seq, please take a look here, http://support.10xgenomics.com/single-cell (more on the data type in Software).

To reiterate, here are some features I envisioned for this project:

Most important

Nice to have

It'd be great if you can think of how to set the visualization up to accommodate these features. Bring your ideas for some brainstorming on Sat morning!

Questions? Thoughts?

See you on Sat! Grace

oganm commented 7 years ago

So as far as I understand there is no real time component as in "we are sequencing this and we need to see the results right now!" sort of way.

gracezheng commented 7 years ago

What I meant by interactive is people can interact with the analysis (eg. query genes, annotate the cells etc).

You can find the dataset I mentioned in my last post here, http://support.10xgenomics.com/transfers/1b77207d5453f23f

Please download the data before the hackathon!

oganm commented 7 years ago

and that is with a pre-existing dataset I assume, we won't be aiming to re-analyse submitted data by third parties

aw18 commented 7 years ago

Hi all, Sorry for joining in on the discussion so late. My name is Albertina and I am currently a Software Engineer at an aerospace company. I don't have an extensive background in bioinformatics but have done some undergraduate research in the field. In my work I use C++ and postgresql, but I am familiar (although it's been a while now) with MATLAB, Java, R, Python and have done a bit of web development using the django framework. I definitely look forward to learning more about bioinformatics at hackseq.

szhan commented 7 years ago

Hi everyone!

Sorry for joining in the discussion so late as well. I am a PhD student in the bioinformatics program at UBC. I study the evolution of polyploid species (organisms with more than two sets of chromosomes) in plants and animals, and I develop phylogenetic methods and math models as part of my research. On the side, I lead the bioinformatics efforts at a startup company, named Fusion Genomics Corp. (http://www.fusiongenomics.com/), that I co-founded during PhD. My company develops NGS-based molecular diagnostic and bioinformatic solutions for infectious diseases. I am comfortable with Perl and R. Looking forward to meeting everyone tomorrow!

Shing