20th Discussion-6 Feb 2018

Attendees

• Brent Richter - Partners
• Alberto Riva - University of Florida
• Alastair Kerr - University of Edinburgh
• Madelaine Gogol - Stowers Institute
• Matt Eldridge - Cambridge
• Simon Andrews - Babraham Institute
• Cory Brauer - UNC Charlotte
• George Bell - Whitehead Institute
• Mark Berres from UW Madison
• Yury Bukhman - Morgridge Institute for Research, Madison WI
• Andrew
• Russell Hamilton - University of Cambridge

Please add yourself...

Training

• How much does your core do, and how do you schedule it?
• What do you teach?
• Do you have any requirements to entry?
• What are your favorite platforms for teaching or teaching tips?

How much training do you do?

• 2-3 full days modeled on software carpentry, done a few times, command line, RNA-seq
• Simon Andrews, Babraham Institute
  • huge amount of training, 50 courses a year, a day each
  • one of their core functions
  • internal / external / commercial venture
  • 5 regular teachers
• Matt Eldridge, Cambridge
  • Also a large amount of training
  • Part of a larger university program
  • demand is massive, long waiting list
  • perhaps have noticed routine analysis going down in volume
    • maybe that's fine, let routine stuff go and focus on harder analyses
• How much? free?
  • 50 pounds a day for postdocs, Phd students free
  • 178 pounds a day with a "day rate"

Curriculum?

• command line, common bioinfo tools
• RNA-seq / ChIP-seq / application specific
• someone suggested github / docker
• Simon said they do
  • statistics
  • core skills - unix, perl/python, R
  • RNA-seq / chIP-seq / methyl-seq
  • "creating figures"
  • gene set enrichment analysis
• Goblet was mentioned
  • but direct reuse is harder than it seems
• Environments are tricky
  • how do you ensure the same environment for everyone?
  • amazon - but then dependent on network
  • some people use docker - last part of course is getting things running on *your* laptop
  • Use a script to check that everything is installed and ready to go

Nanopore

• How is your core supporting Nanopore sequencing?
• Which tools have you found useful thus far for analysis?
• Any particular quality control approaches you are using?

Nanopore

• Who is doing it and what are your experiences?
• Now they have a direct RNA sequencing kit
• minoTour - realtime nanopore monitoring
• Mark, UW Madison - quality of flowcells drops considerally, they run 32-36 hours
• Nanopore is operator / library sensitive
• air bubbles in flowcell can happen
• EDTA will kill it
• Minknow is the software from oxford nanopore, mac/linux/windows, downstream albacore
• smaller flowcells are coming - running a smidgeon on a minion
• minion_qc on github useful QC plots, minimap2 for alignment
• epitome from ONT
• WIMP - what's in my pot
• how about base modifications?
  • nanopolish can report methylated bases, but some question about accuracy. Paper in nature methods
  • Simon mentioned trying to use it to assemble immunoglobulin, found nanopolish was over collapsing their reads

Other New Protocols or Approaches?

• Anything new happening at your institution that you would like to discuss?

Upcoming technologies?

• FPGA / GPU
  • GPU will win out, maybe
  • Intel is pushing FPGA
  • Brent says mostly GPU for their machine learning / AI applications

• Phil Ewels has a github site with tools to examine Novaseq
• Simon mentioned qcfail.com article about novaseq
• Also 10xqc.com for uploading/sharing cellranger reports, nice!

ISMB Workshop, 2018 chicago

• Hiring / interviewing strategies
• Containers / clouds
• lightning talks, unconferencing small group discussions
• We'll be looking for speakers
• Let us know if you want to help organize it!