Difference between revisions of "HiSeq2000 - Next Level Hacking"

From Hackteria Wiki
Jump to: navigation, search
(Links and Information)
(Links and Information)
Line 87: Line 87:
  
 
=== Links and Information ===
 
=== Links and Information ===
<br>
+
 
<br>
 
 
Illumina Next Level Sequencing:<br>
 
Illumina Next Level Sequencing:<br>
 
https://www.youtube.com/watch?v=womKfikWlxM&feature=youtu.be
 
https://www.youtube.com/watch?v=womKfikWlxM&feature=youtu.be

Revision as of 09:37, 9 February 2018

HiSeq.jpg

We got a HiSeq 2000, Next Level Sequencing Machine from the Genomics Facility of Department of Biosystems Science and Engineering in Basel. Contact through Biozentrum, University of Basel. We got it for free with the only disclaimer: "The biohackers should understand that they are responsible to organize and pay for the transport as well as that there is no warranty or support that can be given neither by us nor the DBSSE."

This type of machine seems to be quite difficult to get up and running and also reagents, flowcell-kits and software licences can be expensive. Since more of these machines seem to show up in second hand (there are new machine generations by Illumina) it would be worth trying to find a way to make them work. Sequencing for all.

Specifications:
https://www.illumina.com/documents/products/datasheets/datasheet_hiseq2000.pdf
The HiSeq2000 (200Gb) was introduces in the year 2010. Followed by HiSeq2500 (500Gb) in 2012. And HiSeq X Ten (1000Gb) in 2014. In 2017 the NovaSeq series of machines was launched.
The machine is a quite early on, from March 2011, Serial Number is 700792, so the machine can not be updated to software and chemistry v4. Only machines with SN# 7001403 or higher can get the FPGA update v4.
27747171 1313614108739672 1344364391 o.jpg

First Inspection

I made a first inspection on the machine. It seems very well made (2011). I still think it would be cool to make it run as is. It's basically a big microfluidic system. So if we get the pumps and the cameras to work we can hack it into anything we want 🙂. Even if it's not for sequencing - it's basically a holder for flow-cells with a fluorescence camera attached to it. And 32 channels with pumps and selector valves that attach to the flow cells. Plus a fridge and a computer. And peltier for heating and cooling (pcr). Now trying to get the control software. I also think the system is "relatively" open... the software can be downloaded and kind of installs, there is no ID checking on the supplies or anything. Looks very hackable. Also all the cases can be opened easily. Let's do a weekend hack-session on it.

I think it's great opportunity to learn about next level sequencing and about how theses machines work.

Fluidic System:
27356260 1310155019085581 6425750036040119481 o.jpg

Some pictures from the inside of the machine:

Lausanne Bio-Hackerspace Hackuarium got a HiSeq2000 (SN# 700918) and Gustavo dissected it. Here some pictures that Rachel sent me with comments form what I think components are:

HiSeqDisection.jpgHiSeqDisection2.jpg

Chemistry

Some images describing the Illumina Next-Generation Sequencing Chemistry:
Chemistry.png

Illumina uses a process called "Sequencing-by-synthesis"
The HiSeq (and MiSeq) use 4-colour SBS


The full DNA to Data solution:
Process.png

Paper on the 4 color SBS:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1702316/

Video with the possibilities for library preparation:
https://www.youtube.com/watch?v=_yC0Bzw3WbQ
Library prep kit with sample purification by magnetic beads:
https://www.youtube.com/watch?v=UE1TAZZPUFI

The CBot 2 System is used to prepare the clusters on the flow cell:
https://emea.illumina.com/products/by-type/accessory-products/cbot.html

HiSeq (and TrueSeq) Rapid Cluster Kits can be clustered on the HiSeq (on-board cluster). They also seem to have a lower read and are cheaper:
https://emea.illumina.com/products/by-type/sequencing-kits/cluster-gen-sequencing-reagents/hiseq-rapid-cluster-kit-pe-sr.html

Question:

  • Are the "HiSeq Rapid Cluster Kit v2" compatible with HiSeq (before the v4 update). Or are the "TrueSeq Rapid Cluster Kits"?
  • What software will run on our machine? HCS 2.0.12 or HCS 1.5 ? (depends on software on the FGPA, how to find out?)

Links and Information

Illumina Next Level Sequencing:
https://www.youtube.com/watch?v=womKfikWlxM&feature=youtu.be
Expert Videos:
https://www.youtube.com/playlist?list=PLKRu7cmBQlai-GUWeAN-eHD5xRcCXDW-D
HiSeq2000 support page:
https://support.illumina.com/sequencing/sequencing_instruments/hiseq_2000.html
HiSeq Compatibility Chart:
https://support.illumina.com/content/dam/illumina-support/help/version_compatibility/Default.htm

Video of the scanning:
https://www.youtube.com/watch?v=tuD-ST5B3QA

Discussion


The goal is to make it work!
Let's discuss on the forum.. http://forum.hackteria.org/t/hiseq2000-next-level-hacking/325/1

- Where to get/buy the reagents and flow cells
- Hackquarium Lausanne got a similar machine and dissasembled it. We can get pointers from Gustavo on how to take it all apart!
- Muffatto there's a fluorescence microscope inside (afaik), the issue is to reduce it in size and still have it working
- Muffatto: erik from biocurious reverse engineered the chemistry of the system for BGI - Bengt: Absolutely in on building an openseq kind of thing that can run original reagents. And then re-engineering the system for smaller/cheaper/simpler - even better if combined with an effort to make open reagents - but that is 2 tracks that can progress independently

Technical Descriptions / Findings

The fluorescent readout system with lasers and CCD cameras


HiSeq+Optics.jpg836px-FluorescenceFilters 2008-09-28.svg.png
Patent by Illumina:
https://www.google.com/patents/DE202011003570U1?cl=it

The HiSeq uses an epifluorescence microscope design shown in the diagram. Light of the excitation wavelength is focused on the specimen through the objective lens and the fluorescence emitted by the specimen is focused back the detector by the same objective.

Here you can see the two camera units with even the letters A G T C written on it.
IMG 20180201 221528.jpg

The laser calibration sheets that came with the machine:



The readout system of the HiSeq uses Line Imaging:
Imaging-methods-in-Illumina-instruments-1024x364.jpg

The Flow-Cell


FlowCell.jpg Flowcell.jpg



Some rough calculations/estimations on what is going on in the machine


The DNA clusters are about 1 Micrometer in size. (or bigger for older machines/software, maybe 2-5 um)
The lines on the flow cells are about 1 mm large what means 1000 clusters. (Or lager, up to 1.7 mm)
The reading speed is about 1mm per second or 1000x1000 clusters per second. Or 1 Mega bases.
The lines are 6 cm long and there are 10 lines per flow cell.And two flow cells.
The means 600 Mega Clusters (Bases) per flowcell per complet run. And that takes about 600 seconds or 10 minutes.
Then flush the flow-cell to add the next base (SBS, sequencing by sythesis). Then start over again.
Unitl the whole 150 bases long DNA sequences are read.
This takes about 4 days... 10 Genoms.
The cluster are red in 4 colors / letters at the same time through 2 lasers exciting 4 colors in fluorescence.
By 4 CCD line cameras with Time delay and integration (TDI). Line cameras with 1 micrometer resolution, 1000 lines a second. Or 1 picture 1000x1000 per second...
The chemistry cost some hundreds to some thousands... but for what it does its not so bad. And you get the chemistry kit with the flowcell. So all the magic and the rest is just some kind of state of the art open hardware 🙂

Hardware components

Nice description on what hardware components are used in the HiSeq on the following page (see comments):
https://blogs.swarthmore.edu/Illumina+GAIIx+Teardown/?p=125

  • two lasers (Laserquantum ignis 660nm, gem 532nm with SMD6000 drivers)
  • filter revolvers, beam expanders (Linos 2-8x) followed by a barrel lenses for each wavelength, a combiner to join the two excitation wavelengths
  • piezo actuator for the Z-Stage (Physik Instrumente P-601 with driver E-601 and E-801 Sensor module)
  • Nikon CFI Plan APO VC 20x Objective.
  • XY-table (Parker 803-4099, something like the XR400 series, driven by a ViX-250-IH driver module).
  • The “Docking Station” with the Flowcells are mounted on three Stepper adjustable points to align them with the focal plane of the line scanner.
  • The fluorescence signal is divided by a fixed filter set 4x
  • 4 CCD cameras with S10405 line CCDs from Hamamatsu (DIL 40 package).
  • Two Hamamatsu Camera Control Boards (Model C10000-509) are each controlling two of these line cameras.
  • Illumina board with: line CCDs -> 8 LTC2203 25Msps 16-Bit ADCs -> Altera Cyclone II FPGA. Spartan XC3S4000 FPGA and XC95288 CPLD (both Xilinx)
  • Data is collected by the Phoenix AS-PHX-D48CL Frame grabber card in the Computer

Port usage

Device Port
ARM9BoardSerialPort Port: ARM9CHEM ;CM00006
ARM9BoardDiagSerialPort Port: ARM9DIAG
FPGA Port: IL000004 / 115200 (command_com_port_num

ExcitationFilter1

FPGACommandSuffix = 1
ExcitationFilter2 FPGACommandSuffix = 2
ChromaticCompensator FPGACommandSuffix = 3
TiltMotor_FPGATilt
ZMotor_FPGAControlledPIMotor_250um
ZADCR command

FlowcellFluidics1
FtdiViciValve1: Port: VICIA1 ;CM00004
FtdiViciValve2: Port: VICIA2 ;CM00043
KloehnControllerPump Port: KLOEHNA ;CM00001

FlowcellFluidics2
FtdiViciValve1: Port: VICIB1 ;CM00002
FtdiViciValve2: Port: VICIB2 ;CM00044
KloehnControllerPump Port: KLOEHNB ;CM00003

Laser1,Green532,Smd6000

Port: IL000006
Laser2,Red660,Smd6000 Port: IL000007

XMotor_MDrive_5mm

Port: IL000001
YMotor_VixServoIH_10nm Port: IL000002

Barcode_Reader

Port: FTE2V9ML ; COM5 (com_port_num = 4)
PTC Port: COM2 (com_port_num = 1)/9600
rs232 Port: COM10 (offset 9)
PCIO board ("Test Port") Port: PCIO (com_port_num)

From Video:
Physical Port on Computer Port: COM1
HiSeq Cameras (Phoenix) Ports: COM3 COM4
Virtual Connections Ports: COM5 to COM23

Scanner.ChemistryModule

Drives

Data (D:) Data (E:) DVD (F:) Removable (G:) DoNoEject (H:)

OpenSeq

How about making an Open Source Next Level Sequencing machine - OpenSeq.
Maybe a bit slower and smaller.. like 1 genome per day 🙂
Not sure a price tag of 500k euro is justified for such a machine...
Maybe similar to the iSeq100:
https://www.illumina.com/systems/sequencing-platforms/iseq.html