Post on 20-Jun-2020
SequencingModule 3: Overview
2
Sequencing Workflow
Sample
Preparation
Cluster
Generation Data AnalysisSequencing
3
► The goal of this step is to capture images of the
sequenced DNA to allow the sequence to be
determined
– Given clonal clusters, incorporate fluorescent
nucleotides and take images
► This can be repeated for many cycles
– Incorporate fluorescent nucleotide
– Image tiles
– Cleave terminator
► Paired-end runs require turnaround chemistry similar
to cluster generation
– This happens on the Genome Analyzer, with the
Paired-End Module
Sequencing
Sequencing
4
Basic Sequencing Workflow
Pre-Run Instrument Wash
Clean and Install Prism
Clean and Install Flow Cell
Apply Oil
First-Base Incorporation and Auto Calibration
Steps for completing the
sequencing run vary, depending
on the type of sequencing you are
performing.
Check Quality Metrics
Continue Sequencing Run
Post-Run Instrument Wash
5
Add 4 Fl-
NTP’s +
Polymerase
Incorporated
Fl-NTP is
imaged
Terminator and
fluorescent dye are
cleaved from the Fl-
NTP
X 36 - 100
Sequencing
6
Oil
Flow cell
Prism
Water CavityChiller
Obj.
lens
Camera
Tile
Genome Analyzer Imaging System
7
Autofocus Laser System
► Independent laser system used to calibrate the Z-height
adjustments needed during the run.
► First imaging action performed at start of recipe.
► A series of 30 images are taken, used to create a standard curve.
Movement of AF spot in x-ordinate
x
y
zMovement of objective
8
Autofocus Laser System
The x-position of the
AF spot moves as
the AF images are
taken at varying z-
positions of the
objective lens
► Each image taken at increments of 1000nm in the Z direction
► When starting a run the results are saved in the run folder :
D:\Runs\...\calibration\AFCalResult.txt
9
Autofocus Laser System
► Goodness of fit
► Ideally R(z, r) ~1
► Extraction of spots in x,y coord.
► S(q) value
► Ideally be less than 0.2
► Standard deviation of R values
► Sensitivity
► Ideally: 350 – 400 nm/pixel
► X crossing value of best fit line
X
Y
R
Q
10
A flow cell contains eight lanesLane 1
Lane 8
.
.
.
Column 1
Column 2
Tile
Each lane contains two columns of tiles
Each column contains 50 (GAII) or 60 tiles (GAIIX)
Each tile is imaged four times per cycle –
one image per base.
Flow Cell Images
11
100 Million Clusters
Per Flow Cell
20 Microns
100 Microns
Genome Analyzer Imaging
12
► Sequencing Control Software (SCS) software controls GA and
takes and stores images
► Real Time Analysis (RTA) software analyzes images
– Identifies clusters
– Determines intensities of clusters
– Calls bases for each cluster
– Assigns base call quality scores
– Handles data transfer to Pipeline server
► RTA carries out first two steps of data analysis
Sequencing Software
13
Read length 36 bp 50 bp 75 bp 100 bp
Number of
clusters130-170M clusters
Gigabases
single read /
run time
4.7 – 6.1 Gb /
2.2 days
6.5 – 8.5 Gb /
3 days
9.75 – 12.75
Gb / 4.5 days
13 – 17 Gb /
6 days
Gigabases
paired-end
reads / run time
9.4 – 12.2 Gb
/ 4.3 days
13 – 17 Gb /
6 days
19.5 - 25.5
Gb / 9 days
26 – 34 Gb /
12 days
Avg. raw
accuracy99.25% 99% 98.5% 98.0%
% reads with
no errors> 90% > 80% > 70% >60
GAIIx Performance Metrics
14
Paired-End Sequencing
► Provides long range information
► Important for many short read applications
– Repeat sequences
– Characterize copy number variants & rearrangements
– De novo assembly
– Di-tag sequence cDNAs, ChIP, etc.
– BAC-end sequencing
► Sample multiplexing (identifier tags)
► Increases output per flowcell
15
Valve Flowcell8-way
pump
SBS reagents
1 2 3 4 5 6 7
VICI
8
9 21
SBS reagents
Genome Analyzer
PE module
Priming
pumpwaste
waste
Paired-End Sequencing
Connected
to GA
16
► Sequenced strand
is denatured at the
end of the first read
► 3’-ends of template
strands and lawn
primers are
unblocked
Paired-End Sequencing
Blocked
3’-ends
Sequenced
strand
17
► Single-stranded
template loops
over to form a
bridge by
hybridizing with a
lawn primer
► 3’-ends of lawn
primer is extended
Paired-End Sequencing
Bridge
formation
3’ extension
18
► Single-stranded
template loops
over to form a
bridge by
hybridizing with a
lawn primer
► 3’-ends of lawn
primer is extended
Paired-End Sequencing
Double
stranded
DNA
19
► Bridges are
linearized and the
original forward
template is cleaved
off
Paired-End Sequencing
Double
stranded
DNA
Blocked
3’-ends
20
► Free 3’ ends of the
reverse template
and lawn primers
are blocked to
prevent unwanted
DNA priming
Paired-End Sequencing
Blocked
3’-ends
Reverse
strand
template
21
Add 4 Fl-
NTP’s +
Polymerase
Incorporated
Fl-NTP is
imaged
Terminator and
fluorescent dye
are cleaved from
the Fl-NTP
X 36 - 100
Hybridize
read 2
sequencing
primer
Sequencing Reverse Strand
22
► The GA is the only platform that can do both
– Short Insert Paired Ends
– Long Insert Paired Ends
► Need both for discovery of genome variation
Y
300b
p
Short Insert Long Insert
A A
X
FragmentLigate
(circularize)
Capture, End-
Repair, Ligate
adapters, PCR
X
Y
XFragment, End repair with biotin-NTP YX
X YX Y
Short and Long Insert Paired Ends
23
► Store and prepare reagents as recommended
– 6 month shelf life
– Pool all reagents for a read (2 full kits for 76 cycles)
– Chill reagents before loading on sequencer
► Control ambient temperature
– Ambient temperature specification is 22 +/- 3 degrees
– Temperatures over 28 (or fluctuations) can be seen in intensities
► Monitor first base report and establish a threshold for intensity
► Confirm auto-calibration and monitor autofocus laser performance
► Maintain sufficient network throughput (10 Mbit/second file transfer)
► Ensure that Genome Analyzer is washed regularly
Best Practices for Sequencing
SequencingBest Practices from The Broad Institute
25
Broad Sequencing Workflow
► We usually perform Linearization/Blocking/Primer Annealing (LBPA)
as a single recipe, before sequencing, the day after Cluster
Generation and SYBR QC.
Read 1 Prep
and
Sequencing
Read 2 Prep
and
SequencingAnalysis
Cluster
Gen. &
SYBR QC
Sample
Prep.
26
Broad Sequencing Workflow
Pre-Run Sequencer NaOH Wash
Linearization, Blocking, and Primer Anneal
(LBPA)
Read I Sequencing Reagent Preparation
First Base Incorporation and Calibration
Check Quality Metrics
Prism & Flow Cell Cleaning and Installation
Oil Application
Read I Start & Progression
LBPA
Read 2 Cluster Resynthesis
and Read 2 LBPARead II
Ga2x Post-Run Wash
Cluster Generation SYBR QC
Analysis
Read I
Sequencer Preparation
27
Broad Sequencing Workflow
Read II Sequencing Reagent Preparation
Paired-End Module Preparation
Read II Start & Progression
First Base Incorporation
LBPA
Read 2 Cluster Resynthesis
and Read 2 LBPARead II
Ga2x Post-Run Wash
Cluster Generation SYBR QC
Analysis
Read I
28
► As a beta testing site for Illumina, running various versions of the technology in testing,
limited production and full production, tracking is critical.
► Important points to track from run set-up through post-run analysis:
– Flow Cell ID
– Libraries contained per lane
– Cycle count
– Indexed chemistry
– Recipe version (Stored off-rig, selected at run startup using integrated recipe selector)
– Reagents/FC version (Selected at SCS software start up, tracked in LIMS)
– Lot numbers of reagents (LIMS and paper tracking sheets)
– Software version (Selected at SCS software start up, tracked in LIMS)
Tracking During Run Set-Up
Paper and LIMS tracking
29
Checking Calibration Results
► Standard recipe: Focus Calibration and Edge Find start automatically
after First Base Incorporation chemistry
Recipe change:
► Add User Waits before “Calibration” and “Edge Find”
– Allows for manual course focusing and “four corners check” for oil
– Allows user to monitor Focus Calibration and Edge Find
– Allows user to assess calibration reports
– Allows user intervention (necessary if auto calibration fails)
– Convenient point in recipe to restart if necessary
First Base Incorporation and Calibration
30
Checking Calibration Results
Spec Read 1
R(z, r) Above 0.99
S(q) Below 0.25
MxBrt 1700-2200
Sensitivity 350-400
FQ* Above 65
*See BestZ Plot.
If Max Bright is out of spec, it requires a
service call to adjust the autofocus laser.
Repeating issue may signify dying AF laser.
Illumina Calibration Specs
AutoFocus Z-height Calibration
31
BestZ Plot Shows Focus Quality
Checking Calibration Results
► FQ is also reported in 1st Base
Report (for each base)
► Identifies tile’s highest Focus
Quality value
► If recalibration is necessary (e.g.
lane too sparse, too dense), switch
to a different lane/tile for each
attempt to reduce tile bleaching
► FQ values are dependent on
cluster uniformity and density
► Typical: Lane 4, Column 1, Tile 30
32
188,103 189,265 194,003 183,802
6,983 9,054 3,690 7,773
1,063 1,220 1,128 1,093
73 104 53 97
996 1,087 1,000 971
44 75 51 89
1,379 1,666 1,537 1,496
202 141 263 253
1,190 1,397 1,330 1,184
153 116 220 205
88 87 88 86
2 4 1 5
87 84 86 84
1 4 1 4
88 87 87 85
3 3 4 6
84 83 83 80
2 5 5 7
-229 400 500 -240
7,459 8,059 8,400 8,050
7,688 7,659 7,900 8,290Flow cell Tilt 8,559 9,339 11,240 12,249
Foc Pos Min -750 -2,130 -3,840 -5,340
Foc Pos Max 7,809 7,209 7,400 6,909
T Focus Metric 84 83 86 85
Standard Dev 2 2 2 3
G Focus Metric 88 87 90 89
Standard Dev 2 3 1 2
C Focus Metric 86 85 86 87
Standard Dev 1 2 1 2
A Focus Metric 88 87 89 89
Standard Dev 1 2 1 1
T Intensity 1,416 1,329 1,413 1,396
Standard Dev 114 128 112 186
G Intensity 1,759 1,645 1,778 1,714
Standard Dev 151 224 178 218
C Intensity 1,094 1,008 1,059 1,082
Standard Dev 27 52 50 165
A Intensity 1,216 1,132 1,226 1,202
Standard Dev 69 34 51 142
# of Clusters 191,614 188,073 191,788 180,913
Standard Dev 4,872 4,830 4,749 12,726
First Cycle (1)
MetricName Lane1 Lane2 Lane3 Lane4 Lane5 Lane6 Lane7 Lane8
Machine
Run Date: ReadPrep1
Run Id:
First Base Report: Overview
Calibration and First Base Incorporation
► 1st base reports show
average metrics for each
lane, based on image
analysis of a subset of
tiles
► Generally the images are
taken using a different
exposure than standard
cycle imaging
► Values are similar to what
will be seen during the
run, but not the same
33
188,103 189,265 194,003 183,802
6,983 9,054 3,690 7,773
1,063 1,220 1,128 1,093
73 104 53 97
996 1,087 1,000 971
44 75 51 89
1,379 1,666 1,537 1,496
202 141 263 253
1,190 1,397 1,330 1,184
153 116 220 205
88 87 88 86
2 4 1 5
87 84 86 84
1 4 1 4
88 87 87 85
3 3 4 6
84 83 83 80
2 5 5 7
-229 400 500 -240
7,459 8,059 8,400 8,050
7,688 7,659 7,900 8,290Flow cell Tilt 8,559 9,339 11,240 12,249
Foc Pos Min -750 -2,130 -3,840 -5,340
Foc Pos Max 7,809 7,209 7,400 6,909
T Focus Metric 84 83 86 85
Standard Dev 2 2 2 3
G Focus Metric 88 87 90 89
Standard Dev 2 3 1 2
C Focus Metric 86 85 86 87
Standard Dev 1 2 1 2
A Focus Metric 88 87 89 89
Standard Dev 1 2 1 1
T Intensity 1,416 1,329 1,413 1,396
Standard Dev 114 128 112 186
G Intensity 1,759 1,645 1,778 1,714
Standard Dev 151 224 178 218
C Intensity 1,094 1,008 1,059 1,082
Standard Dev 27 52 50 165
A Intensity 1,216 1,132 1,226 1,202
Standard Dev 69 34 51 142
# of Clusters 191,614 188,073 191,788 180,913
Standard Dev 4,872 4,830 4,749 12,726
First Cycle (1)
MetricName Lane1 Lane2 Lane3 Lane4 Lane5 Lane6 Lane7 Lane8
Machine
Run Date: ReadPrep1
Run Id:First Base Report: Focus
Calibration and First Base Incorporation
► Focal Quality is dependent on
library insert size and cluster
uniformity. Larger inserts will
produce larger clusters
(reducing FQ)
Read 1 Specs:
► Focus Quality Metric > 65
34
First Base Report: # of Clusters
Calibration and First Base Incorporation
► First Base Report does not
always find the same number
of clusters as are found
during run (which uses two
cycles to locate clusters), but
the range should be similar
Read 1 Specs:
► # Clusters: 160-220K
– SCS v2.5, Pipeline v1.5
► # Clusters: 265-320K
– SCS v2.6, Pipeline v1.6
This FC: 188K @ 2.5/1.5
188,103 189,265 194,003 183,802
6,983 9,054 3,690 7,773
1,063 1,220 1,128 1,093
73 104 53 97
996 1,087 1,000 971
44 75 51 89
1,379 1,666 1,537 1,496
202 141 263 253
1,190 1,397 1,330 1,184
153 116 220 205
88 87 88 86
2 4 1 5
87 84 86 84
1 4 1 4
88 87 87 85
3 3 4 6
84 83 83 80
2 5 5 7
-229 400 500 -240
7,459 8,059 8,400 8,050
7,688 7,659 7,900 8,290Flow cell Tilt 8,559 9,339 11,240 12,249
Foc Pos Min -750 -2,130 -3,840 -5,340
Foc Pos Max 7,809 7,209 7,400 6,909
T Focus Metric 84 83 86 85
Standard Dev 2 2 2 3
G Focus Metric 88 87 90 89
Standard Dev 2 3 1 2
C Focus Metric 86 85 86 87
Standard Dev 1 2 1 2
A Focus Metric 88 87 89 89
Standard Dev 1 2 1 1
T Intensity 1,416 1,329 1,413 1,396
Standard Dev 114 128 112 186
G Intensity 1,759 1,645 1,778 1,714
Standard Dev 151 224 178 218
C Intensity 1,094 1,008 1,059 1,082
Standard Dev 27 52 50 165
A Intensity 1,216 1,132 1,226 1,202
Standard Dev 69 34 51 142
# of Clusters 191,614 188,073 191,788 180,913
Standard Dev 4,872 4,830 4,749 12,726
First Cycle (1)
MetricName Lane1 Lane2 Lane3 Lane4 Lane5 Lane6 Lane7 Lane8
Machine
Run Date: ReadPrep1
Run Id:
35
First Base Report: T Intensity
188,103 189,265 194,003 183,802
6,983 9,054 3,690 7,773
1,063 1,220 1,128 1,093
73 104 53 97
996 1,087 1,000 971
44 75 51 89
1,379 1,666 1,537 1,496
202 141 263 253
1,190 1,397 1,330 1,184
153 116 220 205
88 87 88 86
2 4 1 5
87 84 86 84
1 4 1 4
88 87 87 85
3 3 4 6
84 83 83 80
2 5 5 7
-229 400 500 -240
7,459 8,059 8,400 8,050
7,688 7,659 7,900 8,290Flow cell Tilt 8,559 9,339 11,240 12,249
Foc Pos Min -750 -2,130 -3,840 -5,340
Foc Pos Max 7,809 7,209 7,400 6,909
T Focus Metric 84 83 86 85
Standard Dev 2 2 2 3
G Focus Metric 88 87 90 89
Standard Dev 2 3 1 2
C Focus Metric 86 85 86 87
Standard Dev 1 2 1 2
A Focus Metric 88 87 89 89
Standard Dev 1 2 1 1
T Intensity 1,416 1,329 1,413 1,396
Standard Dev 114 128 112 186
G Intensity 1,759 1,645 1,778 1,714
Standard Dev 151 224 178 218
C Intensity 1,094 1,008 1,059 1,082
Standard Dev 27 52 50 165
A Intensity 1,216 1,132 1,226 1,202
Standard Dev 69 34 51 142
# of Clusters 191,614 188,073 191,788 180,913
Standard Dev 4,872 4,830 4,749 12,726
First Cycle (1)
MetricName Lane1 Lane2 Lane3 Lane4 Lane5 Lane6 Lane7 Lane8
Machine
Run Date: ReadPrep1
Run Id:
► Used as QC metric. Desire
intensities as high as possible.
► Library dependent (high GC,
monotemplates, libraries
containing synthetic fillers, etc.
alter T intensity)
Read 1 Specs: Starting
Intensity
Calibration and First Base Incorporation
#
Cycles
Illumina
Specs
Broad
Specs
36 100 400
76 300 600
101 525 1000
126 525 1000
36
Read 2 First Base Report: % Regeneration
188,698 189,035 193,445 188,839
7,001 5,634 5,911 3,070
729 820 770 815
33 43 50 27
699 762 697 739
31 24 36 24
910 1,111 1,069 1,177
45 158 173 159
809 970 934 984
28 122 133 129
86 86 87 86
2 2 2 1
84 84 84 83
1 1 1 1
84 85 86 86
2 3 3 2
80 80 82 81
2 2 4 3
9,160 8,820 8,140 7,020
15,790 15,869 15,769 14,740
6,630 7,049 7,629 7,720Flow cell Tilt 7,940 9,159 10,819 11,960
Foc Pos Min 5,960 4,010 2,000 130
Foc Pos Max 13,900 13,169 12,819 12,090
T Focus Metric 81 80 81 83
Standard Dev 3 3 4 3
G Focus Metric 86 85 86 86
Standard Dev 2 2 2 2
C Focus Metric 83 83 84 84
Standard Dev 1 1 1 1
A Focus Metric 86 86 86 86
Standard Dev 1 1 1 1
T Intensity 1,028 956 965 1,084
Standard Dev 151 160 165 200
G Intensity 1,204 1,139 1,134 1,263
Standard Dev 204 213 209 240
C Intensity 761 714 709 833
Standard Dev 35 42 46 184
A Intensity 848 795 793 896
Standard Dev 46 57 57 157
# of Clusters 191,374 189,536 191,547 188,875
Standard Dev 2,831 3,905 3,740 3,580
First Cycle (1)
MetricName Lane1 Lane2 Lane3 Lane4 Lane5 Lane6 Lane7 Lane8
Machine
Run Date: ReadPrep2
Run Id:
► % Regeneration is calculated
using average T intensity
across all lanes from R1 & R2
(Read1 T int) / (Read2 T int) x 100
Read 2 Guidelines:
► T Regeneration: > 50%*
► * Lower than 50% is fine if
Read 1 had particularly high T
intensities
This FC: 75.67%
(7765/10655)x100%
Read 2 First Base Incorporation
Fail Modes
38
Tracking Fail Modes
► For each failed run, a Fail Report is
created to track relevant information
► Assist in troubleshooting future runs
► Identify trends, bin issues (e.g. by date)
Flowcell Fail Report
(Double click on boxes to check or uncheck)
Run Name/ FC: 42KF5AAXX
Date of Failure: 9/30/09
Sequencer/CS: XAP
Reported by: SG
Reviewed by: SC
Read 1 Read 2 Single Read Other
Description of Failure Mode: CS issue – describe
Reagent top off issue
Low Intensity at 1st Base = 614.13
Poor Re-synthesis for Read 2
Machine Crash due to Lack of Memory
Known Mechanical Issue:
Analysis issue:
Other Issue (please include as much detail as possible):
Action Taken: Run Cancelled
Please update Squid for all run cancellations
Run Failed Analysis: Please update Squid for all run failures
Unresolved : If unresolved, please list the people aware of the issue:
Other:
Please include the following FC details:
Comments/Description: Run had low t int. to begin with, cancelled b/c got worse as run
continued, would lead to a bad 126cycle run. Rehyb failed causing peeled matrix.
Libraries and loading concentrations: Solexa-14942, 41, 40, 39, 38, 37, 36, 35.
WR ID: 20066
39
Tracking Fail Modes
► Most common on-sequencer fail modes in the past 6 months:
– Oil propagation
– Low intensities at 1st base of read I and II
– RTA processing issues
– Mechanical/hardware problems
Fail Mode July –December 09
40
Pareto: Fail modes
Nov 09-Jan 10July –December 09
Major fail modes change over time
41
Fail Mode: Oil
Worse in later cycles; Oil expands when heated, wicks onto FC surface, then spreads
► Portions of, or the entire tile is out of focus
► Identify oil issues as early as possible, as FC can be cleaned and run resumed
► Images, and RTA FQ charts are key tools in identification
42
Fail Mode: Tile Out of Focus
► A distorted focus dot usually precedes a tile that is out of focus
► Possible cause: Oil or reagent (leaking from manifold) on surface of flow cell
► Possible cause: Problem with focus laser
► Possible cause: RTA cannot process the tile due to extreme density
43
Fail Mode: Oil on Side of Prism
► Oil on the optical side of the prism
► Optical pathway is compromised
reducing transmission of laser light
through the prism and into the FC.
Not here
Oil goes here
Not here
44
Fail Mode: OilIdentify/Resolve/Prevent (Best Practices)
► Identify
– RTA charts
– Images
► Resolve
– At run prep: remove assembly (FC & Prism) and clean and reinstall
– During run: stop run, clean FC, prism, & peltier with methanol, restart run
► Prevent
– Use conservative, standardized oiling method (to be presented in lab module)
– Clean peltier block frequently
45
Fail Mode: RTA Processing “Thread Death”
► RTA has 3 processing threads that
are responsible for tile processing.
► If a thread dies, a “checkerboard” tile
pattern will show on the charts.
► Depending on the number of threads
that died, the corresponding number
of tiles will be skipped.
46
Fail Mode: RTA Processing “Thread Death”Identify/Resolve/Prevent (Best Practices)
► Identify
– RTA charts, tile status
► Resolve
– If identified during Read 1, stopping and restarting RTA using the restart file in the
run folder will usually restart the processing threads that died.
– If identified during Read 2 after RTA has completed processing Read 1 with the
same phenotype, RTA will not be able to reprocess the lost tiles.
► Prevent
– Monitor Read 1 closely after starting, looking for the “checkerboard” tile processing
47
Fail Mode: Poor Linearization
► Can occur at either read 1 or 2
► Characterized by platelet- or donut-
shaped clusters
► Low flow of LMX on cluster station
or PEM is an indicator
Read 1 & 2 Start & Progression
48
(block)
Not all bridges
get linearized
Only edge of
cluster fluoresces
(block)
Good Linearization
Poor Linearization
What we think is happening
Cluster in bridge form
Primer only
incorporated
on linearized
strands
Normal Linearization Efficient Primer
incorporation Nucleotides incorporated
only on primers
49
Fail Mode: Poor Linearization Identify/Resolve/Prevent (Best Practices)
► Identify
– Insufficient flow on the CS or PEM is the primary indicator
Volume checks at LBPA, especially of critical reagent LMX
– Inefficient LMX reagent
Images on sequencer
► Resolve
– Strip primer (denature with NaOH) and repeat LBPA
► Prevent
– Mark reagent level on each tube
– Avoid freeze-thaw of reagents
– Avoid storage temperature fluctuations
– Record Lot #s
50
Fail Mode: Poor Blocking
► Characterized by few visible
clusters and a very bright
background
► Can be subtle (slight increase in
background noise)
► Unblocked OH groups at the ends
of linearized strands can also
incorporate fluorophores, resulting
in increased background noise
51
…but not all of
them get
blocked.
Incorporate onto
cluster AND
unblocked primers
Poor Blocking
Good Blocking
Excess P7 primers on
FC have –OH group,
What we think is happening
Excess P7 primers
on FC have –OH
group
Cluster in bridge form Normal Blocking
All excess primers
blocked
Nucleotides incorporated
only on primers
Normal Primer
incorporation
Normal Primer
incorporation
52
Fail Mode: Poor BlockingIdentify/Resolve/Prevent (Best Practices)
► Identify
– Insufficient flow on the CS or PEM is the primary indicator
Volume checks at LBPA, especially of critical reagent BMX
– Inefficient BMX reagent
Images on sequencer
► Resolve
– Strip primer (denature with NaOH) and repeat LBPA
► Prevent
– Mark reagent level on each tube
– Avoid freeze-thaw of reagents
– Avoid storage temperature fluctuations
It’s common to see a combination of inefficient Linearization & Blocking on one FC
53
Fail Mode: Poor Primer Annealing
► Inefficient primer annealing is
characterized by a lack of cluster
intensity. Clusters are visible but the
intensity is significantly below the
minimum threshold.
► In this case, the Primer Annealing
step can be repeated (a “rescue
rehyb”)
54
“Rescue Rehyb”
► If 1st Base Report shows low T intensity, the Primer Annealing step can
be repeated (a “rescue rehyb”)
► Rescue Rehyb:
– Denature primer by flowing NaOH
– Wash
– Repeat primer annealing
► Enables us to save both Read 1 & Read 2
► Success rate:
– 50% over all attempts
– 100% over all instances identified certainly as poor primer incorporation
► Can be performed without removing flow cell from sequencer, using the
Paired-End Module, but this requires disconnection of tubing
► Illumina is developing Rehyb recipe and kits for CBot
55
110,611 112,584 115,076 121,981
4,751 8,056 5,478 5,779
51 72 314 146
12 34 93 91
56 75 296 140
12 33 76 80
209 307 1,201 630
63 163 369 392
93 130 465 253
26 62 134 151
T Intensity 216 229 529 768
Standard Dev 122 82 233 166
G Intensity 456 508 1,136 1,743
Standard Dev 263 213 555 350
C Intensity 124 139 320 523
Standard Dev 67 52 154 104
A Intensity 105 119 271 437
Standard Dev 58 48 135 94
# of Clusters 127,178 127,496 140,946 141,146
Standard Dev 8,154 4,457 13,454 14,493
First Cycle (1)
MetricName Lane1 Lane2 Lane3 Lane4 Lane5 Lane6 Lane7 Lane8
131,723 132,009 131,675 131,873
1,431 651 660 971
469 505 497 500
17 16 29 21
427 458 447 446
19 12 22 17
2,466 2,662 2,642 2,672
131 112 219 167
974 1,045 1,029 1,027
49 46 69 53
T Intensity 1,287 1,244 1,308 1,279
Standard Dev 113 137 118 109
G Intensity 2,734 2,691 2,652 2,447
Standard Dev 140 146 187 277
C Intensity 579 556 560 547
Standard Dev 25 45 26 32
A Intensity 507 498 491 467
Standard Dev 19 21 32 36
# of Clusters 131,471 133,575 138,406 139,278
Standard Dev 2,696 3,646 8,123 8,602
First Cycle (1)
MetricName Lane1 Lane2 Lane3 Lane4 Lane5 Lane6 Lane7 Lane8
Avg T = 311
Avg T = 1149
“Rescue Rehyb”
After Rehyb:
1st Base Report
Initial Primer
Hyb:
1st Base Report
56
“Rescue Rehyb”
After Rescue
Rehyb:
Avg T = 1149
Initial Primer
Hyb:
Avg T = 311
57
Fail Mode: Poor Primer AnnealingIdentify/Resolve/Prevent (Best Practices)
► Identify
– Low 1st base intensity
► Resolve
– Perform “Rescue Rehyb”
► Prevent
– Monitor reagent consumption and flow
– Mix primer tube thoroughly after thawing
58
Fail Mode: Uneven Primer Lawn
Possible defects in the substrate used for primer seeding process in flow cell manufacturing
Tracking when and where (cycle/tile/run) these occur can distinguish between defective substrate and
problems with the laser /optics assembly (signified by reproduced pattern in subsequent runs)
59
Fail Mode: Switched Reagent
► Incorporation Mix where Scan Mix
should be
► Unattached fluorescent nucleotides
were not flushed away with Scan Mix
► Labeling bottles and color matching
labels to avoid switching reagent
bottles
60
Fail Mode: Mode Scrambler Failure
► The mode scrambler is a device that
spreads the laser illumination
uniformly over the tile.
► Identify
– Mottled image surface
► Resolve
– Service engineer
intervention required
– Run can continue but
expect higher error rates,
phasing and prephasing
values, low % PF
► Prevent
– Replace mode scrambler
61
Fail Mode: Footprint
► A “footprint” is caused by a misaligned beam shaper is dark band along any side of
the tile (it is often confused with oil edges)
► User corrective measure will be demonstrated in the lab section
62
Fail Mode: Torn/Delaminating Substrate
► Substrate peeling from the FC
surface
► Sometimes occurs after
multiple rescue Rehyb
attempts
► Sometimes occurs when
Cluster Station or Sequencer
Peltier block fails
Best Practices
64
Best PracticesReagent Handling
► Use 175ml Falcon & 150ml Corning bottles to eliminate top-offs in runs >100b
► Label and color-match bottles to lines
► Perform NaOH wash before every run start and before prolonged down-time
► Measure and track flow volumes: dead volumes and waste volumes
► Mark levels of reagent volumes when loading on instrument
► Filter Incorporation mix (IMX) to reduce “super clusters”
– Filter buffer & nucleotides before adding enzyme (enzyme won’t filter)
► Control reagent temperatures and thawing environment
– Thaw each tube of reagent upright in room temp water
– Immediately transfer to ice, once thawed
– Change gloves after handling cleavage components
– Thaw and store separately from other components
– Cleavage components must not contact other reagents
65
Best PracticesStage/FC/Prism Assembly
► Clean stage with methanol before every flow cell is loaded
– Use water on ports to avoid degradation of manifold gaskets
► Clean peltier with methanol
► Use lens paper to clean optical surfaces such as prism and flow cell
► Clean flow cell with water, then methanol
– Avoid ports to prevent DNA degradation
► Check for proper seal between flow cell and manifold by flowing buffer
– A steady line of bubbles indicates improper seating
► Oiling
– Oil flow cell slowly from the left side to avoid dripping oil onto optical edge of prism
– Ensure oil flows under the entire flowcell
– Put a 10ul tip on top of 200ul tip to help guide oil under flow cell
– Drag a clean tip along the right edge of flow cell can help to seal oil underneath
– Check images in four flow cell corners for adequate oil
(Will be covered in the lab module)
66
Best PracticesRun Monitoring
► RTA (covered in Working with Data Module)
– Identify freezes and processing
errors
– Ensure sufficient intensity in all
channels
– Ensure focus quality
– Identify reagent leaks and oil issues
– Identify sudden changes
► Images
– Ensure focus
– Identify “super clusters,” especially
near end of reads
– Identify reagent leaks and oil issues
67
Key Points
► Careful run monitoring and tracking are keys to successful run
► Recognizing fail mode phenotypes allows early identification of problems,
sometimes before they become detrimental to a run
► The technology is not “set and forget”; It requires user activity throughout
the process
► Each user/group must determine an appropriate set of rules and
expectations concerning tracking, QC, metrics, fail points, etc. based on
downstream effect on data quality
Image Gallery:
Unique Phenotypes
69
Images: Unique Phenotypes
Frosted objective Dried substrate
70
Images: Unique Phenotypes
► Crop circles?
71
Images: Unique Phenotypes
► “Ocelot”
– Jen Fostel, Hayward
72
Images: Unique Phenotypes
► Amoebas? (SYBR QC image)
73
What We’ll See in the Lab
► Loading the GA
– Cleaning flow cell & prism
– Loading & oiling flow cell
– Quality Checks during flow cell loading
► GA maintenance & troubleshooting
– Washes
– Flow rates
– Footprints
– Max brightness
– Focus laser
► Best practices