day5, jackson lab

==========
Mark Adams
microbial genomics service

mock microbial community => assess DNA extraction method, or other procedures


==========
Aditya Srikanth Kovuri
Sandeep Namburi

NIST, cloud characteristics,  
http://csrc.nist.gov/publications/nistpubs/800-145/SP800-145.pdf
on-demand self-service
broad network access
resource pooling
rapid elasticity, https://www.techopedia.com/definition/29526/rapid-elasticity
measured service

https://biologydirect.biomedcentral.com/articles/10.1186/1745-6150-7-43

Amazon S3

glalaxy cloudman

Google cloud is cheaper than GoogleCloud.
GoogleGenomics API.

https://cloud.google.com/genomics/reference/rest/

Dockers


Microsoft Azure Research awards

Google Research award

day4, jackson lab,

=> Krish Karuturi
big data genomics, computational and informatics challenges

https://www.jax.org/research-and-faculty/tools/scientific-research-services/computational-sciences/staff/krishna-karuturi

https://github.com/TheJacksonLaboratory/civet
TORQUE resource manager

https://en.wikipedia.org/wiki/Comparison_of_cluster_software

benchmarking pipelines

https://en.wikipedia.org/wiki/SNV_calling_from_NGS_data#List_of_available_software

https://www.nature.com/articles/srep43169

GSEA
GSA, Effron & Tibshirani

XENOME

etherpad
https://public.etherpad-mozilla.org/p/2017-05-18-bigData-grad-prof


===============================
Peter Robinson, Ph.D., The Jackson Laboratory for Genomic Medicine
Phenotype driven genome analysis

https://scholar.google.com/citations?user=TPOD_XUAAAAJ&hl=en

Ontology, disambuilgous terms.

human phenotype ontology

information content (IC) of concept.

semantically similar diseases scores

PhenoBLAST
Washington NL 2009, Plos Biology

======================================
Y Ada Zhan, ChIP-seq

https://en.wikipedia.org/wiki/ChIP-sequencing

encodeproject.org
https://academic.oup.com/bib/article/17/6/953/2453197/A-comprehensive-comparison-of-tools-for

bd2kuser@ip-172-31-73-47:~/ChIPseq$ cat readme.txt
###################
# ChIP-seq module #
###################

# ChIP-seq data
In the directory ChIPseq/
 GM12878_control_chr1.fastq
 GM12878_CTCF_chr1.fastq

# Genome
In the directory ChIPseq/hg38/
 GRCh38.chr1.fa
 GRCh38.chr1.size

# Tools
 fastqc (quality check)
 bowtie (sequence mapping or alignments)
 samtools (manipulating alignments in SAM format. BAM format is a compressed version of SAM file)
 macs2 (peak calling)
 bedtools (to handle sequence coordinate files in BED format)
bd2kuser@ip-172-31-73-47:~/ChIPseq$




bd2kuser@ip-172-31-73-47:~/ChIPseq$ cat workflow.sh
# quality check
fastqc GM12878_control_chr1.fastq
fastqc GM12878_CTCF_chr1.fastq

# Prepare genome
bowtie-build hg38/GRCh38.chr1.fa hg38/GRCh38.chr1

# Mapping
bowtie -m 1 -S ./hg38/GRCh38.chr1 GM12878_control_chr1.fastq > GM12878_control_chr1.sam
bowtie -m 1 -S ./hg38/GRCh38.chr1 GM12878_CTCF_chr1.fastq > GM12878_CTCF_chr1.sam

# Further processing
## compress to BAM
samtools view -bSo GM12878_control_chr1.bam GM12878_control_chr1.sam
samtools view -bSo GM12878_CTCF_chr1.bam GM12878_CTCF_chr1.sam
## sort
samtools sort GM12878_control_chr1.bam GM12878_control_chr1.sorted
samtools sort GM12878_CTCF_chr1.bam GM12878_CTCF_chr1.sorted
## index
samtools index GM12878_control_chr1.sorted.bam
samtools index GM12878_CTCF_chr1.sorted.bam

# Peak calling
macs2 callpeak -t GM12878_CTCF_chr1.sorted.bam -c GM12878_control_chr1.sorted.bam -f BAM -g 175000000 -n GM12878_CTCF_chr1 -B -q 0.01

# Check the peak model
Rscript GM12878_CTCF_chr1_model.r

# Motif analysis
## extend summits 100bp on both directions
bedtools slop -i GM12878_CTCF_chr1_summits.bed -g hg38/GRCh38.chr1.size -b 100 > GM12878_CTCF_chr1_summits_ext.bed
## get sequence file (i.e. fasta)
bedtools getfasta -fi hg38/GRCh38.chr1.fa -bed  GM12878_CTCF_chr1_summits_ext.bed -fo GM12878_CTCF_chr1_summits_ext.fa
## The .fa file will be uploaded to MEME online server for motif discovery (http://meme-suite.org/tools/meme)


BED file format


MEME motif discovery

ChiPseek website for interactive data analysis,

day3, 20170517Wed Jackson Lab, Galaxy, IGV,

=> Paola Vera-Licona
gene network

time series gene expression data -> network

structure-based control of signaling networks (optimization of interaction? )

HER2-positive breast cancer

BiNoM           -> geneXplain --> OCSANA
gene expression -> list TFs ---> mapping pathways + master regulator --> identify optimal combination of intervention from network analysis

candidate genes with p-values
pick largest connected component
using random sampling permutation to evaluate the choice of p-value cutoff.

https://binom.curie.fr/
http://compsysmed.org/Software/OCSANA/OCSANA.html

Using annotated pathway to build a directed nework for intervention analysis and prediction.

How drugble? Drug reposition?

Q: KEGG?

==================================
=> Reinhard Laubenbacher

https://www.ncbi.nlm.nih.gov/myncbi/browse/collection/46337356/?sort=date&direction=descending

http://www.sciencedirect.com/science/article/pii/S1040842813002308

================
Karl Broman, Reproducible research (should added to my REU bootcamp training).

biostatistics and medical informatics

http://kbroman.org/

https://github.com/QinLab/Talk_ReproRes

http://kbroman.org/steps2rr/


IGV: need *bam file for alignment, *bai file for index. 

vcf file can be visualized in IGV or Ensembl Variant Effect Predictor.

http://www.cbioportal.org/

Usually, large genes tend to have more mutations than small genes. Genes with repetitive elements tend to have more mutations.

genomespace.org

day 2, afternoon, 20170515 jackson lab

genome data sources
https://repositive.io/product/datasources/

genomes in a bottle
https://discover.repositive.io/datasets/c1802ea2-a853-476d-91e6-c08290cc6fe4

Carl Zimmer

George Church
http://www.personalgenomes.org/


JAX HPC 256G RAM per node, 20 cores per node,

day2, moring, 20170516

=> Sheng Li, RNAseq
RNAseq library contruction

Kukurba KR, montgomery SB, Cold Spring Harbo Protoc, 2015,
https://www.ncbi.nlm.nih.gov/pubmed/25870306

For microRNA, ~20nt, special protocol is required.

stranded and non-stranded library (to distinguish overlapping exons or genes on opposite DNA strands)

minimal reads: 20-25 millions reads  for mammalian transcriptiome

Illumina Hiseq-4000, ~ 4000 millions per lane. 4-8 libraries per lane. Often, double indexing can be used for high number of multiplexing libraries.


2nd step, Gene annotation: GenCode gencode-help@sanger.ac.uk
Ensembl88,
GTF format

3rd step, gene expression quantification

RNAseq metric,
single-end RPKM, reads per kilobase per million reads
paired-end, FPKM, fragments per kilobase per million reads
nomalize read counts for sequencing depth, length of gene
TPM, transipts per million
 pro: sum of total normalized reads is the same for all samples.(not for  R/FPKM)

before 1st step, Quality check step.
 genebody coverage, (with genes)
 insert sizes
 GC content
 reads distribution
 adaptor enrichment (containmination or PCR amplification bias?)
 read quality

RSeQC, Liguo Wang, Bioinformatics 2012
FastQC
  polyA selected 3' UTR, so 5'UTR degradation can be a problem.

Public data:
GEO
RNA-seq blog

http://rpubs.com/shelly1436/274304

combatR, correct of batch effect

biological degradation of mRNA during aging, using sva latent variable, to distinguish biological degradation from non-biological degradation.

https://gist.github.com/slowkow/6e34ccb4d1311b8fe62e#file-rpkm_versus_tpm-r


=======================
Single cell RNAseq, Ion Mandoiu

psuedotemporal order of cells
https://www.nature.com/nbt/journal/v32/n4/pdf/nbt.2859.pdf

single cell mutaional profieing and clonal phylogeny in cancer
Potter, Genome Re
http://genome.cshlp.org/content/23/12/2115.long

cell type identification in primary visual cortex


Fluidigm
https://www.ncbi.nlm.nih.gov/gds/?term=fluidigm

https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE77477

challenges in single-cell RNAseq: low RT and sequencing depth, "zero inflated" data

DAVID
GeneMania

Matching clusters to cell types or organism parts

10X genomics . https://www.10xgenomics.com/ .
neuron cortex
https://support.10xgenomics.com/single-cell-gene-expression/datasets

https://support.10xgenomics.com/single-cell-gene-expression/datasets/1M_neurons

Jackson lab genomics Day 1

Moring

install anaconda3 on ubuntu virtualbox

#
hqin@rainboxdash:~/anaconda3/bin/$ ./jupter notebook &


https://thejacksonlaboratory.github.io/bd2k-workshop/ 

samtools view example.bam | less

Linux exercises: up and down arrows, tab for file-names autocomplete

bam file

vcf file (variant call file?)

less test.vcf

mkdir tmpdir
cd tmpdir
cd ..
ls

Essential probability and statistics for introduction to big data

• summary stattistics vs empirical statistics
• Common data transformation, Z-scores
• Bayesian inference
• Multiple hypothesis testing

https://en.wikipedia.org/wiki/Median_absolute_deviation

Bayes's rule

Introductory Data Mining

=====================
Afternoon

curricula:
 all biology are computational
 computational biology are parasite of biology

Approach:
  research project approach using real datasets
  targeted students: professional and academic oriented?

data carpentry's R for genomcis
http://www.datacarpentry.org/R-genomics/


RStudio projects
https://support.rstudio.com/hc/en-us/articles/200526207-Using-Projects

args(barplot)
?lm
??lm
help.search("kruskal")

==========
Barke Southern Illinois Medical School, longest-living mouse

===========
Brenton Gravely, RNA genomics

Dscam, over 100 exons, an extreme case of alternative splicing
Schmucker 2000 Cell. Mutually exclusive splicing.
Ig Repeats.  Dimerization are isoform specific.
https://www.ncbi.nlm.nih.gov/pubmed/19934230

Dscam variatiosn between species
Gravely 2005, Cell.

Competing RNA base-pairing is a common mechanism for mutually exclusive splicing in anthropods, Yang 2011 Nature Struc Mol Biol
http://www.nature.com/nsmb/journal/v18/n2/abs/nsmb.1959.html

single cell RNA sequencing
Drosophila S2 cells, each cell show the same splicing isoform.

Drop-seq of Drosophila and human cell to control the number of cells in each droplet.

Oxford nanopore sequencing

1500 RNA binding proteins in human genome

Van Nostrand Nature methods, 2016, eCLIP-seq reveqls RBP-specific binding profiles
http://www.nature.com/nmeth/journal/v13/n6/abs/nmeth.3810.html