NOTE

Report
Servers, R and Wild Mice
Robert William Davies
Feb 5, 2014
Overview
• 1 - How to be a good server citizen
• 2 – Some useful tricks in R (including ESS)
• 3 – Data processing – my wildmice pipeline
1 - How to be a good server citizen
• Three basic things
– CPU usage
• cat /proc/cpuinfo
• top
• htop
– RAM (memory) usage
• top
• htop
– Disk IO and space
• iostat
• df -h
Cat /proc/cpuinfo
[email protected]:~$
processor
:
vendor_id
:
cpu family
:
model
:
model name
:
stepping
:
microcode
:
cpu MHz
:
cache size
:
physical id
:
cat /proc/cpuinfo | head
0
AuthenticAMD
21
2
AMD Opteron(tm) Processor 6344
0
0x600081c
1400.000
2048 KB
0
[email protected]:~$ cat /proc/cpuinfo | grep processor | wc -l
48
Memory can be in RAM or in
swap
Htop and top
Swap is BAD
Ideal use – 0
(in this case it is probably residual)
RAM – 512GB total
142 in use (rest free)
Load average – average over
1, 5, 15 minutes
48 cores
[email protected]:~$ iostat -m -x 2
disk use
Relatively unused
High sequential reading (fast!)
Also note state – D = no IO
There are also ways to optimize disk use for different IO
requirements on a server – ask Winni
Get sizes of directories
Disk usage
Get available disk space
How to be a good server citizen
Take away
• CPU usage
– Different servers different philosophies
– At minimum, try for load <=number of cores
• RAM
– High memory jobs can take down a server very easily and will
make others mad at you – best to avoid
• Disk IO
– For IO bound jobs you often get better combined throughput
from running one or a few jobs than many in parallel
– Also don’t forget to try and avoid clogging up disks
• P.s.
– A good server uptime is 100%
2 – Some useful tricks in R (including
ESS)
• R is a commonly used programming language /
statistical environment
• Pros
– (Almost) everyone uses it
– Very easy to use
• Cons
– It’s slow
– It can’t do X
• But! R can be faster, and it might be able to do X!
Here I’ll show a few tricks
ESS (Emacs speaks statistics)
• Emacs is a general purpose text editor
• Lots of programs exist for using R
• There exists an extension to emacs called ESS
allowing you to use R within emacs
• This allows you to analyze data on a server
very nicely using a split screen environment
and keyboard shortcuts to run your code
I have code on the left
An R terminal on the right
Running a line of code ctrl-c ctr-j
Running a paragraph of code ctrl-c ctrl-p
Switching windows ctrl-x o
Google: ESS cheat sheet
http://ess.r-project.org/refcard.pdf
C- = ctrl
M- = option key
R – mclapply
• lapply – apply a function to members of a list
• mclapply – do it multicore!
• Note there exists a spawning cost depending
on memory of current R job
Not 19X faster due to
chromosome size differences
Also I ran this on a 48 core
server with a load of 40
/dev/shm
• (This might be an Ubuntu Linux thing?)
• Not really an R thing but can be quite useful for R
multicore
• /dev/shm uses RAM but operates like disk for
many input / output things
• Example: You loop on 2,000 elements, each of
which generates an object of size 10Mb. You can
pass that all back at once to R (very slow) or write
to /dev/shm and read the files back in (faster)
ff
• Using save, load is an easy way to interact
with data most of time time. ff allows you to
use variables like a pointer to RAM
Bonus – you can use mclapply to different entries without collisions! (same entries =
collisions)
Rcpp
Some things R is not very good at –
like for loops
I write the c++
code as a text
vector
You can call
fancy R things
from c++
Simple example of c++ use in R: for a
reference genome (say 60,000,000),
coded as integer 0 to 3, determine
the number of each possible Kmer of
size K (ignoring converses for now)
I then “compile” it here (takes ~1-3
seconds for simple (<1000) line
things)
I often just pass in
lists, pass out lists
(Note that upon making this slide I realized there is
an easy way to do this using vectors)
table(4^0 * ref[seq(1,refT-K)] + 4^1 * ref[seq(2,refTK+1)] + … ) and adjusting for NAs
Complicated example I actually use
• Want to do an EM HMM with 2000 subjects and
up to 500,000 SNPs with 30 rounds of updating
• Using R – pretty fast, fairly memory efficient
– Use Rcpp to make c++ forward backward code in R
– for iteration from 1 to 30
• mclapply on 2000 subjects
– Run using Rcpp code, write output to /dev/shm/
• Aggregate results from /dev/shm to get new parameters
– Write output (huge) to ff for easy downstream
analysis
2 – Some useful tricks in R (including
ESS) – Take away
• A lot of people complain about R being slow but it’s
really not that slow
– Many packages such as Rcpp, ff, multicore, etc, let your
code run much faster
– Also, vector or matrix based R is pretty much as fast as
anything
• If 1 month after starting to use R you are still copying
and pasting, stop what you’re doing and take a day to
learn ESS or something similar
– If you don’t use R often you can probably ignore this
– (P.s. I copied and pasted for 2 or 3 years before using ESS)
3 – Data processing – my wildmice
pipeline
• We have data on 69 mice
• Primary goals of this study
– Recombination
• Build rate maps for different subspecies
• Find motifs
– Population genetics
• Relatedness, history, variation
• Admixture
Caroli
N=1 - 40X – WDIS = Wild derived inbred strain
Famulus
N=1 - 40X - WDIS
F
WildSpret
N=1 - 40X - WDIS
W
WildDom
N=1 - 40X - WDIS
W
W
Migration
France
M. m. Domesticus
N=20 - 10X - Wild
C
weight
Classical
0.5
M. m. musculus
M. m. Castaneus
Taiwan
0.2
N=1 - 40X - WDIS
N=10 - 30X - Wild
WildCast
10 s.e.
0.1
N=13- 40X – Lab strains
WildMus
India
0
0.0
W
N=1 - 40X - WDIS
N=20 - 10X - Wild
0.3
0.4
6 pops – 20 French, 20 Taiwan,
10 Indian, 17 Strains, 1 Fam, 1
Caroli
bwa aln –q 10
Stampy –bamkeepgoodreads
Add Read group info
Merge into library level BAM using
picard MergeSamFiles
Picard markDuplicates
Merge into sample level BAM
Use GATK RealignerTargetCreator on each population
Realign using GATK IndelRealigner per BAM
Use GATK UnifedGenotyper on each population to create a list of putative
variant sites.
GATK BaseRecalibrator to generate recalibration tables per mouse
GaTK PrintReads to apply recalibration
69 analysis ready BAMS!
Example for 1 mus caroli (~2.5 GB genome ~50X coverage)
Downloaded 95GB of gzipped .sra (15 files)
Turned back into FQs (relatively fast) (30 files)
bwa – about 2 days at 40 AMD cores (86 GB output, 30 files)
Merged 30 -> 15 files (215 GB)
stampy – cluster 3 – about 2-3 days, 1500 jobs (293 GB output, 1500
files)
Merge stampy jobs together, turn into BAMs (220 GB 15 files)
Merge library BAMs together, then remove duplicates per library, then
merge and sort into final BAM (1 output, took about 2 days, 1 AMD)
1BAM, 170 GB
Indel realignment – find intervals – 16 Intel cores, fast (30 mins)
Apply realignment – 1 intel core – slower
1 BAM, 170 GB
BQSR – call putative set of variants – 16 intel cores – (<2 hours)
BQSR – generate recalibration tables – 16 intel cores – 10.2 hours
(note – used relatively new GATK which allows multi-threading for this)
BQSR – output – 1 Intel core – 37.6 hours
1 BAM, 231 GB
NOTE: GATK also has scatter-gather for cluster work – probably
worthwhile to investigate if you’re working on a project with 10T+ data
Wildmice – calling variants
• We made two sets of callsets
– 3 population specific (Indian, French, Taiwanese),
principally for estimating recombination rate
• FP susceptible – prioritize low error at the expense of
sensitivity
– Combined – for pop gen
• We used the GATK to call variants and VQSR to
filter
What is the VQSR? (Variant Quality Score Recalibrator)
model PDF
12
12
HaplotypeScore
HaplotypeScore
10
10
lod
!4
8
!2
6
0
4
2
2
4
8
6
!
filtered
!
retained
4
2
0
0
!2 !1
0
1
2
3
4
!2 !1 0
QD
1
2
3
4
QD
10
8
training
6
4
!
neg
!
pos
HaplotypeScore
Take raw callset. Split into known and novel (array, dbSNP, etc)
Split into known and novel
Fit a 12
Gaussian Mixture Model on QC parameters on known
Keep10the novel that’s close to the GMM, remove if far away
12
HaplotypeScore
outcome
8
6
4
2
2
0
0
!2 !1
0
QD
1
2
3
4
Ti/Tv -> Expect ~2.15 genome wide
Higher in genic regions
novelty
!2 !1 0
QD
1
2
3
4
!
novel
!
known
8 It’s a good idea to benchmark your callsets and decide on the one with the parameters that
suit the needs of your project (like sensitivity (finding everything) vs specificity (being right))
Population
Training
Sensitivity
HetsInHomE
chrXHetE
nSNPs
TiTv
arrayCon
arraySen
French
Array Filtered
95
0.64
1.97
12,957,830
2.20
99.08
94.02
French
Array Filtered
97
0.72
2.28
14,606,149
2.19
99.07
96.01
French
Array Filtered
99
1.12
3.62
17,353,264
2.16
99.06
98.09
French
Array Not Filt
95
2.06
5.82
18,071,593
2.14
99.07
96.58
French
Array Not Filt
97
2.97
8.24
19,369,816
2.10
99.07
98.01
French
French
French
French
French
Array Not Filt
17 Strains
17 Strains
17 Strains
Hard Filters
99
95
97
99
NA
6.11
1.29
2.20
4.19
5.36
15.73
3.89
6.52
11.63
16.37
22,008,978
16,805,717
18,547,713
20,843,679
19,805,592
2.01
2.14
2.11
2.04
2.06
99.06
99.07
99.07
99.06
99.09
99.20
93.49
96.49
98.62
96.96
Sensitivity – You set this – How much of your training set do you want to recover
HetsInHomE – Look at homozygous regions in the mouse – how many hets do you see
chrXHetE – Look at chromosome X in males – how many hets do you see
nSNPs – number of SNPs
TiTv – transition transversion ratio – expect ~2.15 for real, 0.5 for FP
arrayCon – Concordance with array genotypes
arraySen – Sensitivity for polymorphic array sites
We chose a dataset for recombination rate estimation with low error rate but still a good
number of SNPs
Population
Training
Sensitivity
HetsInHomE
chrXHetE
nSNPs
TiTv
arrayCon
arraySen
Taiwan
Array Not Filt
95
2.05
11.20
36,344,063
2.12
NA
NA
Taiwan
Array Not Filt
97
2.87
14.67
39,183,932
2.10
NA
NA
Taiwan
Taiwan
Taiwan
Taiwan
Taiwan
Array Not Filt
17 Strains
17 Strains
17 Strains
Hard Filters
99
95
97
99
NA
6.34
1.83
2.16
3.66
6.11
25.57
10.32
11.20
15.80
19.44
42,864,322
29,748,456
34,112,325
39,549,666
33,692,857
2.05
2.11
2.11
2.08
2.04
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Indian
Array Not Filt
95
1.11
1.80
66,190,390
2.18
NA
NA
Indian
Array Not Filt
97
1.59
2.57
71,134,757
2.16
NA
NA
Indian
Indian
Indian
Indian
Array Not Filt
17 Strains
17 Strains
17 Strains
99
95
97
99
3.70
0.67
1.09
2.63
5.56
1.16
1.63
3.31
78,220,348
57,674,209
65,981,654
75,103,886
2.11
2.18
2.17
2.13
NA
NA
NA
NA
NA
NA
NA
NA
Indian
Hard Filters
NA
5.41
72.61
78,487,616
2.10
NA
NA
All
Array Not Filt
95
1.90
8.95
140,827,810
2.04
99.07
96.74
All
Array Not Filt
97
2.38
13.99
160,447,255
2.03
99.07
98.20
All
Array Not Filt
99
4.52
22.73
184,977,157
1.99
99.06
99.36
Some of the datasets are extremely big
Combined datasets allow us to better evaluate differences between populations
Notes – VQSR sensitivity not always “calibrated” –
Note: Be VERY skeptical of the work of others wrt sensitivity, specificity, that depends on
NGS. Different filtering on different datasets can often explain alot
French and Taiwanese very inbred, not so for the Indian mice
Homozygosity (Blue) for chromosome 19
10
10
6
4
10
0
Taiwan
10
20
30
Position Mbp
Homozygosity = red
Huge Taiwan and French
bottleneck, India OK
20
India
5
Mouse
0
2
15
Mouse
8
20
France
Homozygosity (Red) for chromosome 19
40
50
60
30
Position Mbp
40
50
60
Caroli
WildMus
WildDom
Classical
Recent Admixture is visible in French
and Taiwanese populations
France
Migration
weight
WildCast
0.5
Taiwan
India
0
WildSpret
10 s.e.
0.0
Famulus
0.1
0.2
0.3
0.4
Admixture / introgression common
Our Domesticus hotspots are
enriched in an already
known Domesticus motif
0.4
0.2
0.3
Broad scale correlation is conserved between
subspecies, like in humans vs chimps
0.1
Pearson correlation
0.5
0.6
French hotspots are cold in Taiwan and vice-versa
0
1000
2000
3000
Window size (kb)
4000
5000
Conclusions
• 1 – Don’t crash the server
• 2 – There are tricks to make R faster
• 3 – Sequencing data is big, slow and unwieldy.
But it can tell us a lot
Acknowledgements
•
•
•
•
•
•
Simon Myers – supervisor
Jonathan Flint, Richard Mott – collaborators
Oliver Venn – Recombination work for wild mice
Kiran Garimella – GATK guru
Cai Na – Pre-processing pipeline
Winni Kretzschmar – ESS, and many other things
he does I copy
• Amelie Baud, Binnaz Yalcin, Xiangchao Gan and
many others for the wild mice

similar documents