Preparation: extracting high-quality DNA and RNA
Vânia Costa, Applications Scientist, Oxford Nanopore Technologies
- For cell lines, ONT recommends TRIzol extraction.
- For long-term storage, keep DNA and RNA at -20°C or -80°C
- Flick to mix DNA
- ONT usually uses Megaruptor 3 for shearing DNA (like a syringe; can change speed and alter fragment length)
Q&A
- RNA storage
- we would recommend to extract the RNA and store it at -80°C until ready to use
- The best storage condition for extracted RNA is -80°C. A week at -20°C might not be too detrimental, but it is best to avoid storing it at -20°C for too long.
- DNAse treatment: we frequently treat our RNA extractions with DNAse and we have not observe any degradation
Preparation: selecting the right library prep method for your experiment
Tomasz Dobrzycki, Technical Applications Scientist, Oxford Nanopore Technologies
- Nanopores embedded in an electrically-resistant membrane
- Passing DNA through nanopore creates specific peturbations into the current
- Ligation is optimised to produce the highest output
- Rapid has been further optimised recently for producing ultra-long reads
- Ultra-long with rapid: 10-20 Gb yields on MinION flow cell, still greater than 50kb N50 libraries
- Direct RNA
- Specific RT adapter is initiallly annealed to RNA
- Reverse-transcription (optional) improves sequencing yield by about 30%
- Adapter is slightly different, processes RNA at 100 bases per second
- cDNA sequencing
-
20M cDNA reads from MinION, >160M cDNA reads from PromethION flow cells [with SQK-PCS111 kit]
- Can use bespoke RT adapter to enrich for molecules of interest
-
- 16S barcoding kit
- Quicker, more accurate than other 16S kits
- Adaptive sampling
- Software-level
- Hopefully future uses will enable [barcode?] balancing usage
Choosing The Right Nanopore Sequencing Device For You
Bryant Catano, Technical Applications Scientist, Oxford Nanopore Technologies
- Flongle flow cells
- Can use for pre-QC (load onto Flongle first, then other cells)
- MinION flow cells
- 10-20 Gb ultra-long (>50kb) reads
- PromethION
- 100-200 Gb
- Flongle / MinION
- works in Mk1b / Mk1c / GridION
- MinION Mk1C
- all necessary compute inside the device (+ touchscreen)
- highly portable, field-ready
- GridION Mk1
- higher-end GPU (V100)
- Can run up to 5 individually-addressable flow cells (most flexibility possible)
- Flongle
- TMO: 2.8 Gb [maximum output?]
- PromethION
- P2: 2 individually-addressible; P24 / P48: 24/48 respectively
- P48 up to 14 Tb per run
- A100 acquisition unit
- Adaptive sampling
- Coming soon to PromethION soon
- Genomic epidemiology
- GridION being used to great use for large-scale SARS-CoV-2 viral genome sequencing
Sequencing: how to load a Flongle Flow Cell
Andrada Tomoni, Sequencing Development Scientist, Oxford Nanopore Technologies
- Flongle - latest development in terms of sequencing; one-time use, cheap, disposable flow cell
- Should expect 1-2 Gb data per run
- MinION case
- Metal lip holds both adapter and flow cell
- Need to be careful around posts, slots for adapter pins
- Flongle adapter
- Contact pins where Flongle flow cell will slot in
- Pins on the bottom - make sure they're not bent before inserting
- ASIC chip should be completely covered by a heat pad
- Insertion - rest it loosely under the lip, and let it guide you
- Flongle flow cell
- Make sure no dirt or grease
- Insert in the same way as the adapter; should get a satisfactory click
- Do not disconnect it until the sequencing run is finished
- New Flongle design; cosmetic differences
- Slightly bigger loading port area
- Waste channels a bit clearer
- Use the lid; it ensures proper thermal control
- A little bit of liquid remaining on top of the Flongle flow cell array is fine; it will flow into the array eventually
Q&A
- How do you deal with with bubbles underneath the loading port? What
do you do if there is air in the loading port or pipette tip?
- Hello, in order to avoid any air bubbles underneath the loading port, it is recommended to ensure that there is no air in the pipette tip when priming the flowcell.
- What do you do if the Flongle valve is blocked and resists
pipetting?
- Hello, can you please contact Technical Services via the conference platform? Thank you
- How do you deal with crystallised flow cell storage buffer in the
Flongle loading port?
- Hi, it shouldn't be a problem as long as is not blocking the loading port. Best course of action would be to clean with a lint free wipe.
Analysis: basecalling your data and detecting methylation
- Basecalling
- Raw read accuracy has been steadily increasing over the years
- current Bonito CRF model
- Using new training sets
- Modification basecalling (e.g. 5mC) will include a supplementary table with modified base calling probabilities
- Output as fastq-only is not recommended, as re-calling or alternative analysis is not possible
- Fast5 files contain analyis information
- Command-line tools offer more control over results
- SUP model is 3x slower than HAC, which is 3x slower than Fast model
- Methylation
- R9.4 better reproducibility and lower coverage needed in comparison to bisulphite
- Remora with R10.4 is even better
- Can call and phase long nanopore reads using methylation
- methplotlib - visualisation of per-nucleotide methylation probability
- Working with HTS, JBrowse & IGV to display methylation
- Native long read data enriches datasets, even at low coverage
Welcome to London Calling 2022
Gordon Sanghera
-
[... missed out the start]
-
Translation of sequencing into community is a long process
- Human, environmental, animal, plants
-
Ceremonial gong to formally open LC 2022
Emma White
-
Trying to integrate online and on-site experience; virtual audience displayed in breakout rooms and auditorium
-
Amazing lineup of speakers
The potential of ultra-rapid nanopore genome sequencing for critical care medicine [video]
Euan Ashley, Stanford University, USA
- Picture of Ferrari, used in place of Moore's Law graph
- $0.01 if price had dropped as much as human genome sequencing had dropped
- Have focused on accuracy as well as cost, need great confidence in variants
- Sometimes speed is important as well
- Matthew (13 year old boy)
- Dry cough, assumed he had COVID
- Diagnosed with heart failure
- Heart was barely managing to push out 10% of blood
- In cardiogenic shock - heart not providing enough blood
- Surgical implantation of an artificial heart
- Possible myocarditis (reversible), but also possible a genetic cause
- Normally would do a biopsy of the heart; they did that (it didn't show lymphocytes that were expected for myocarditis)
- 10 years ago, considered rapid sequencing process: 50 hrs (42 hours with process improvement); pushed down to 26 hours (GWR); down to 19.5 hours.
- In 2021, 6-8 weeks is about the amount of time a genome assembly normally takes. Patients like Matthew don't have even 5 days.
- Idea pitched to John Gorzynski - "Would it be possible to do it faster?"
- First call to ONT; really only 1 machine that could do it; P48 1X genome in 2 minutes
- Considered how to do it faster; maximising speed and quality, spreading 8 library preps over the whole PromethION
- Compute was a problem
- Process
- Wet lab 4.2h
- Sequencing 2.7h
- Dry lab 3.6h
- Sampling 62 mins
- fragmentation 19 mins
- Sequencing 1.26 Gb/min
- Compute power at the time PromethION received, expected 80 more hours
- Used a cloud method to optimise
- Basecalling with Guppy GPU
- minimap2 alignment
- Reducing file size every few minutes
- GPU 4x Tesla V100 per instance; 48 CPUs per instance
- Moved from compute overhead of 16 hours to overhead of 30 mins
- Alignment was 3.6h
- Found very quickly a heterozygous duplication that indicated genetic abnormality in TNT gene; had shown it could not be myocarditis
- Total processing time 11.2 hours
- Excited to bring team together, and apply this technology
- Highly variable time on the curation side, depending how long to get variants
- Other optimising
- Pore occupancy was better without barcoding
- no change in accuracy (small amount of carryover, but still wouldn't affect numbers)
- Removing barcoding step reduced genome to all under 12h
- Patients presenting with neurodevelopmental disorders
- Head of facility would like to sequence 2/3 of patients; direct benefit to patients
- costs about $10,000 per day for ICU; test can save the healthcare system money
- Having beaten the GWR, were happy to be awarded a new GWR
- We need to rethink how we're doing genomics if shifting from weeks to minutes and hours
- If made Ferrari as much faster as human genome speed, would change from 199 miles/h to 1/20th speed of light
Q&A
Question hosted by Dan Garalde, who took the first conference question
- What does it take to scale for hospitals around the world?
- Tech is much more scalable than it initially appears
- Will soon be limited by the difference between the patient and the sequencing machine (e.g. sequencing in hospital vs flying sample to a sequencing centre)
- Everything shipped up to the cloud; anyone who wants to set this up could do it almost anywhere
- Last mile (telemedicine) could also be done remotely
- How would you convince insurers and health care providers to take this on?
- They are already convinced
- Moves at California and federal level, and around the world, to specifically fund neonatal and intensive care genome sequencing
- Cost saving; a few thousand dollars now, will drop further, reducing length of ICU stays
- Are there certain genetic conditions that make more sense for this testing?
- Asked physicians first - any patients for who a genetic diagnosis would help
- A lot of neurological conditions (42% diagnostic rate in pilot; higher after implementation)
- Multi-system disorders
- 2/3 of patients could benefit from this technology
Identification of Novel genomic structures and regulation patterns at HPV integration events in cervical cancer
Vanessa Porter, PhD Candidate, University of British Columbia
- HPV necessary, but not sufficient for development of cervical cancer
- HPV infects basal layer through microabrasions in the cervix
- Cancer follows from integration of HPV into genome (70% of cancers), often E2 is disrupted or removed, can lead to increased expression of oncogenes
- Rearrangements can be simple, or more complicated
- In cancer, HPV integrates near oncogenes [image: circular plot, showing bars that stack into the centre]
- Project objective - investigate integration of HPV
- Looking at genomes of 50+ Cervical cancer samples
- Identify integrations
- Assess methylations
- HIV Tumor Molecular Characterization Project
- Sequenced 44
- Performed Nanopore analysis pipeline: basecalling, alignment, SNV, methylation, phasing; added on other custom workflow
- Almost all samples; HPV integrates between regions where at least two double-stranded breaks have occurred - an integration event
- Used hybrid HPV/human genome, allowed to identify integration events as translocations
- Long reads can pair together breakpoints on the same read; can link chromosomes
- Sample reads subsetted to create assemblies; categorised integration events into an integration type
- 3 major resolutions
- Deletion of region
- Duplication of region, get alternating human/HPV structure
- eccDNA (extrachromosomal circular DNA); extra region excised as circular DNA, never see reads from circular region. Distance between breakpoints consistently longer, about 70kb (may need this to be stable in the cell)
-
- eccDNA breakpoints can be simple (one region), or complex (multiple breakpoints)
- Content of inserted HPV sequence can be variable, all contain E6 and E7 (oncohenes), so probably the minimally-required genes
- Ran PCR to confirm circular contigs
-
- Breakpoints on different chromosomes (chr4, chr12)
- Copy number loss at the breakpoints
- Proposed mechanism is an unbalanced translocation event
-
- Multi-breakpoint integration events
- Presents of multi-breakpoints leads to instability
-
- Repeat region integration
- Integration associates with an expansion of repeats
- Some samples have multiple integration events
- HPV16 was more integrated than others
- Some integration event categories are more common
- most common: multi-breakpoint
- Presence of categories can change depending on HPV type
- Using methylation for looking at haplotypes; found clear demethylation upstream of an integration region (100kb, up to 400kb away)
- Looked at HPV genome at sites of integration
- Compared with non-integrated samples
- Some samples have different methylation at different regions
- Integrated HPV is hypermethylated vs unintegrated HPV
Q&A
- How do you assess false positive and false negative results? Phenotype data for correlating events?
- A spike-in chromosome to detect false-positive chimeric reads, high redundancy required (at least 5 reads); probably missing a lot of sub-clonal integrations
- Survival data; found HPV clades associated with survival; are limited by sample size
- How are complex integrations detected?
- Deletions / translocations can't be easily detected; had to get creative
- Have been packaging up the process
- Sniffles for HPV breakpoint detection
Spotlight Session
Chaired by Thomas Bray, Commercial Director Greater China
- 2-minute pitch, followed by vote to see winning talk, presented in the auditorium
- Theme: how sequencing is changing your area of research
The bacterial blackmarket: utilising sequencing to study the inter-cellular
Richard Goodman
- Star wars scroller - The bacterial black market
- Bacteria outwitted
- Mobile colistin resistance gene emerges
- Nanopore sequencing is helping in the fight against the AMR wars
- Bacteria evolve resistance, share through a mechanism analogous to a black market
- Used a decoy sequence to evaluate movement of sequences through the bacterial black market
Whole mitogenome variation in domestic cats discovered via nanopore sequencing
Emily Patterson
- July 2012 - dismembered torso of a man; forensic scientists didn't find human DNA, but found cat hair; prime suspect had a pet cat
- Nanopore sequencing can be used for looking at cat DNA
Unloicking the transcriptomic architecture of bacterial viruses with ONT-cappable-seq
Leena Putzeys
- Viruses - bad diseases & global pandemics
- But... not all virus are bad; some can cure disease
- Bacteriophages, can kill bacteria; have biotechnological potential
- Want to take advantage of nanopore sequencing platform to read out viral sequences
[sleep]
Update from Oxford Nanopore Technologies
Introduced by Rachel Rubinstein, Technical Product Manager, Software
Novel/interesting statements selected by David Eccles
- Smaller new ASIC; 5mm x 6mm, 400 channels, 1 MUX, expecting around 9 Gb from a thumb-sized device, expecting 10 Gb output with deeper wells
- Adaptable sequencing from Kit 14: 400 bases/s at 99.2% modal accuracy (default), 260 bases/s at 99.6%, 520 bases/s at 99.0%
- Longest perfect accuracy read 72 kb
- New base caller - Dorado - coming out this year: open source, user extensible, apple + NVIDIA
- PCR tiling kit for VolTRAX (4 samples), 30 mins HOT, sample to sequencing for about $100
- RAD chemistry will be fully released Q4 2022; looking great, reducing input to about 100ng
- iPad case MinION will replace Mk1B
- Clive Brown's cell-free DNA data is now available
Clive Brown
- Only 300 people at London, over 1000 people online watching, last done in 2019; too much stuff, got department heads to do most of it
- Clive will do a bit at the end; things I think are interesting
- Guys here do pertinent updates (latest changes), a lot of meaty detail not seen before
- Clive gave an update about a month ago, a lot of progress since then
- Online a long line of previous talks
- Our new ability to measure modified bases
- Relationship between speed of sequencing and throughput (explained by Lakmal, interesting progress; what Lakmal tells you is half what it needs to be)
- First James Clarke (chips)
- Each person 15 minutes, 35 slides, so might skip past a few slides for my time
James Clarke
- The Platform - all the things you require to do nanopore sensing - the base:
- Nanopore sensing
- Nanopore sequencing
- Applications
- Your research
- 3 Different types of flow cell: MinION, PromethION, Flongle
- Devices: Mk1C, Mk1B, GridION, PromethION P24/P48/P2
- Got here by continuous improvement and agile development
- Some people want a firmer foundation - Q Line
- Clive gave a talk in March, a move to R10 series
- Ability to go as fast as R9 series (this was holding R10 back a bit)
- Simpler product offering
- R10 output, different ways to run
- Default 400 bases/s, similar to R9 speed, 4 Versions of output
- PCR / E. coli (236 Gb)
- Native human, no interaction 10kb, 166 Gb
- 30kb, 155 Gb, with wash kit interactions
- 80kb N50, 120 Gb
- Matt Loose got 170 Gb from his run
- Default 400 bases/s, similar to R9 speed, 4 Versions of output
- Deeper wells on PromethION are helping (from Marathon Flow cell)
- Like a battery: the more of it you have, more runtime, more run consistency
- PromethION is the best flow cell: most recent ASIC (best accuracy); more pores
- P2 can run 2 PromethION flow cells
- John Tyson has the device
- Out in the live lounge in London
- Lots of unfinished R&D business
- Would like to change flow cell design
- Larger flow cells could go to 4-chamber loading (on PromethION)
- 600 microns of fluid above the nanopore, a bit of a hangover from early-access stage
- Team will completely refactor this
- Want to integrate library prem
- Chemistry changes (e.g. silver chloride)
- Can get accuracy over Q20 with silver
- should go to Flongle first
- Manufacturing
- Nanopore sequencing getting more popular; want quality to remain the same
- Economies of scale - quality, output
- Adding new features via a method of continuous improvement
- Monitoring; characterising problems; making sure you capture that problem
- Where are we going?
- Products enabled by ASIC; now have SmidgION ASIC
- Consumes a lot of power; redesigned with low-end output, but as good as it can be from bottom up
- Proof of concept - working properly; don't want to be eventually in Flongle format
- Got enough data to basecall, looks pretty nice, enough to feel confident to moving to next step
- 3 components for new flow cell; want to keep path to ASIC short, won't add much more to cost
- Have a PCB which will be form factor of chips (about thumb size)
- 5mm x 6mm, 400 channels, 1 MUX, expecting around 9 Gb, expecting 10 Gb output with deeper wells
- Products enabled by ASIC; now have SmidgION ASIC
Lakmal Jayasinghe
- Need to do many different things
- Need to improve speed through pore
- Need to get maximum number of pores
- Last year: R9 + Kit 10
- Raw read accuracy about 98.3
- Introduced new nanopore, R10.4; new motor E8.1; new kit Kit 14
- Raw read accuracies beyond 99% for first time
- Speed initially 250 bases/second
- Relationship between speed and accuracy is complex; important to improve speed and accuracy together (a hard job)
- Want to keep accuracy about 99%
- Teams screened different nanopores, 8000 nanopores, 300 (3000?) motor proteins
- Were also changing run conditions: temperature, buffer conditions
- At the end of the speed spectrum, accuracy was lower
- Picked particular enzyme and nanopore wth 400 bases/s, changed models to push accuracy above Q20
- new pore 10.4.1
- Speed improved, better speed distribution
- E8.2 accuracy has improved, better defined errors and fewer mis-steps
- Picked particular enzyme and nanopore wth 400 bases/s, changed models to push accuracy above Q20
- Other nice thing about new chemistry, can tune using software by adjusting temperature
- 400 bases/s, 99.2% accuracy
- lower temperature to 260 bases/s, 99.6% accuracy
- higher temperature 520 bases/s, more output, 99.0%
- Duplex mode Q30 accuracies
- Released to new developers, most have pleasing experience
- Duplex chemistry
- Working on very hard to implement
- Capture second strand as well, 5% of normal experiments complement strand follows template strand
- Accuracy increases
- Needs perfect sample prep
- Needs ends adapted
- Needs no nicks in DNA
- By June last year, rate improved to 20%
- Now have 40%; target is 70%
- Duplex process can be run on both R9 and R10 pores
- Trying to make sure that kit14 is built around duplex
- Q31 achieved with current duplex caller, going to improve much further, will be available soon
- Improving chemistry / accuracy further
- R10 distance about 5nt between read head
- New pore (Rx) has about 9nt between read heads, talked about at NCM last year
- About 98% accuracy
- Random substitutions to sample prep using PCR methods; breaks homogeneity of homopolymer regions
- Homopolymer length accuracy improves
- Consensus accuracy also improves
- Half-way through? Don't know what we'll do when we reach 100%
Stuart Reid
- Data, where we are, where you can expect to be
- Changing temperature changes other things
- e.g. current increases
- We are not at the optimum condition
- A lot of headroom for ASICs
- Making changes deep in the software
- Have architecture, training
- Several different sorts of accuracy
- Raw read accuracy; interested in single molecules, population diversity
- Duplex, can call 2 reads
- More than 2, consensus, the average of a population
- The few errors that remain in raw reads are mostly averaged out in consensus
- Have SNPs, structural variations, base modifications
- A few talks today showing really high test accuracies
- Where are we now?
- Raw read, focusing on 400 bases per second
- 230 Gb from PromethION flow cell, >Q20
- Run condition choices - accuracy at 260 bases/s; output at 520 bases/s
- Output 307 Gb
- Accuracy 99 Gb (Q31 modal)
- Longest Q30 duplex read 260 kb
- Longest Q40 144 kb
- Longest perfect read 72 kb
- Probably quite a lot of headroom in the algorithms
- Accuracy progression
- Duplex reads have very few errors
- Consensus / 400 bp
- Complete, contiguous, accurate genomes with nanopore data alone
- Recommend nanopore-specific tools
- Optionally polishing with medaka; Flye still showing Q50 bacterial assemblies without Medaka
- Variants - Like Clair3 to call; LRA for variants
- SNPs 99.9% accuracy
- SVs 95% (hard for ground truth)
- Base modifications
- All the base modifications are built into the nanopore signal
- With Remora, built into MinKNOW, coming out in next release; visualisations just work, can see in popular genome browsers
- SV and methylation ground truth is hard to come by
- Claim that nanopore is the most accurate technology for detection
- Highly contiguous, accurate human assemblies
- Duplex Q30 (99.92%)
- Flow cell record now at 307 Gb
- Read length remains at 4.2 Mb
- Software features
- Short-fragment mode now built into the software; nanopore has read any length of DNA; lower length now configurable (down to 20 bases)
- Accuracy is not a function of read length; applies to short reads to
- Adaptive Sampling
- If you don't like what you're sequencing, can kick it out
- 5-10 fold improvement
- Now enabled on PromethION (3X improvement now, beta release coming out)
- Don't panic; most won't need to do anything
- New file format - pod5: pore open data
- Thanks to slow5 format, working around many downstream analysis problems
- Had to focus on the right performance; 48 flow cells on PromethION
- 4-fold improvement in write performance, 10-fold improvement in read performance
- available on github pod5-file-format
- New base caller - Dorado - coming out this year
- Want rapid reach
- Open-source and user extensible
- Will support apple as well as NVIDIA
- Should be live on github
Rosemary Sinclair Dokos
- We have been building together a platform that can perform the full biology for an experiment
- Any read length from 20bp up to millions of bases
- Scalable, deliver data in real time
- Want to scale platform to fit your biology
- Flongle, MinION all the way up to P48
- Made accessible through starter packs
- Partnership with (Evantor?)
- Keeping platform open
- Key features open to tweak: number of nanopores, run time
- Ultra-long reads, multiplex
- Aiming to generate data at really competitive prices
- Seamless upgrade paths
- Most of platform upgrades are contained in flow cell, sequencing kits, software; have done 221 updates in last 3 years
- Latest release, have gone from research to customer in about 3 months
- Rapid training, Cambridge compute has halved time from R&D
- Like to put proof-of-concept out quickly
- Miten: from flow cells to talk in 12 days
- Some people want more stability
- Developer
- Early access
- Released
- Fully-released
- Q-line
- Diagnostics
- Extraction - over 72 protocols; also extraction central on community
- Long-read sample preparation
- Will be releasing short-fragment elimination kit
- Also adding new ultra-long kit, as many extraction methods as possible
- cDNA sequencing (see talk by Dan Turner)
- Barcoding kit (PCS111) now shipping
- Working on size selection for isoform detection
- Single cell protocol
- cDNA lab protocol simplified
- Teams have released sockeye - nanopore-only analysis pipeline
- Direct RNA; a lot of questions
- Need so much input material
- Teams have reduced input requirements from 500ng to 50ng
- COVID-19
- Continue to support community with ARTIC & Midnight kit, sequenced over 1M genomes together
- Faces from LC2015, has been magical to see faces over zoom over last 2.5 years
- As rolled out globally, a light-bulb moment: nanopore sequencing is good for pathogen detection
- Teams actively working with community to provide a sample-to-answer workflow (like ARTIC and Midnight solutions)
- Diagnostics team working on TB and respiratory metagenomics
- Voltrax carried on in background
- PCR tiling kit (4 samples), 30 mins HOT, sample to sequencing for about $100
- Now working on next product pipelines
- Whole genome PCR
- Ligation sequencing
- Want to get VolTRAX into flow cell loading zone
- Kit 14
- The best of all worlds - output from R9 + accuracy from Kit 12
- Key thing: ligation kit, with developers today
- Native barcoding kits very close - 24 / 96
- New kits contain all sequencing components
- Getting through 384 barcodes, in developer's hands (interested people can talk to us)
- Allows genome sequencing at $4 per sample
- Rapid-based preps
- RAD chemistry looking great, reducing input to about 100ng
- Preparing for launch Q3 2022
- An enormous amount to juggle
- Currently in development for Kit 14 for Ligation
- R10.4 and Kit 12 being taken off the store Q4 2022
- Laser focused getting R10.4 and Kit14 into release phase
- Sequencing
- Raising the bar on outputs
- Flongle could be doing 3.4 Gb
- 60 Gb on MinION
- From Clive - needs to be doubled
- They're not short or long reads, they're just reads
- File formats
- Raising the bar on accuracy
- Raising the bar on outputs
- All features making it through to smaller and portable devices
- P2 enables people to start exploring PromethION-style experiments
- Human WGS for under $1000 [assumes purchase of 96 flow cells, and one sample per run]
- Have insitutions with capacity for over 2000 genomes per week
- Sample to answer at under 8 hours
- Human at depth - dark regions and methylome
- Analysis
- Github research releases
- Tutorials - set of packaged tools
- Epi2ME lab workflows
- Epi2ME
- Simple interface to Epi2ME, no command line
- Anyone interested in replicating can look online
- Building a full picture
Clive Brown - One Little Thing
- A complicated story
- MinION Mk1D - We are making this thing, an iPad case, wrapping a MinION/Flongle
- Get 5G, cloud streaming
- Low power, chance of running for 6h is quite high
- Can register now
- Will replace Mk1b
- Module at the side to embed miniature VolTRAX
- Accuracy - nanopore basecalls have high confidence; errors tend to be confined to homopolymers
- Ability to call modifications; almost impossible not to in nanopore; a single atom change
- Currently blank-over modified bases
- A huge amount of [research] calling modified bases by default
- Problem is ground truth; empirically-testable with synthetic sequences
- Basecaller, now synthesising DNA
- Nanopore has a program to make a synthesiser
- Can now make a synthetic set of random bases with 15th base mC, often sequences you wouldn't see in nature
- Random training sets now outperforming bisulphite training sets
- Proof of concept looks pretty good
- First data in 2012, nobody remembers, based on outy method
- We run inny now: motor on front of the molecule acts as a break
- Can suck in DNA, motor drags DNA out of pore
- Has some interesting features
- Looking at scheme for re-reading molecule, more than one scan on a given molecule
- Average single-molecule accuracy pretty much where we were on inny 18 months ago, no reason why it can't be improved further
- Average overall accuracy is moving quite quickly
- Per-base calibrated confidence scores can be moved quite high
- Dwell-time distribution of steps is different, so should be okay for resampling
- Asked to sequence lambda without amplification; they took cell-free DNA and measured a mutation at high confidence with a background - "Needle in a haystack"
- Results are so good that they believe they can see PCR artefacts
- I get bled every so often, the substrate for nanopore research
- Extract DNA, fractionated
- Bottom red wine fraction; top white wine fraction - an old haematologist joke
- Sequenced my DNA, can call methylation patterns, can be correlated with tissue of origin, can sequence modified cell-free DNA
- See patterns that match expectations
- With nanopore, can see longer cell-free fragments
- Most platforms have a cap
- Is the longer stuff newer?
- A whole number of applications for short fragments
- Looks quantitative, can call to same accuracy, can call modifications
- Cell-free run available; ultra-long Kit14 version of me will be available as well, can play with those
- Would like to add a duplex version of me
- Will ultra-long cell-free catch on?
- David Cameron attended a flow cell loading class - everyone jams pipette in the hole
- New loading assistance
- Have a little insert that fits into the loading hole
- can squirt sample into the pan, or tip from Eppendorf, can put sample directly in
- Quite sophisticated
- Interested in liquid biopsy, sample prep on the flow cell, everything occurs on the flow cell
- Top chamber for swab with buffer
- Barrier with a number of things including gel
- Electrophoretically move DNA from swab to input chamber
- Motors that will grab DNA, flow cell comes with chemistry
- Outy sequencing system, from end-to-end swab or spit of blood
- Proof-of-concept
- Button on the flow cell can be pushed to flush the flow cell
- All geared up towards moving the device out of the lab
- Vision of field sequencing
- Think of ideal device, make a chemistry that works with the device
- Cell-free DNA, short fragments, sequencing multiply, brought together into a decentralised liquid biopsy, all in one device over the next year or two
- Early blood glucose, gave 20mL blood to get blood glucose measurement, moved onto strip tests
- Can now put device on your arm with needle that measures glucose
- Toothbrush - the ideal sampling device; twice a day, looking at 10-20 ml blood over a month
- If see a molecule that looks suspicious, sequence it again, and again
- Clive's fantasy product - trace over time; with infection / cold, accumulate enough to do something about it
- Protein sequencing
- Very active development
- Not the only chemistry we're looking at, but this is the one I'm talking about today
- Can put a molecule through a nanopore as long as it's charged
- Existing technology based on mass spectrometry, results difficult to interpret, quite posterior loaded
- Think that same platform for DNA and RNA can be used for protein
- Double-stranded DNA/RNA
- use DNA motor for moving protein - conjugate to DNA
- Generate signal; see DNA signal, then peptide signal
- A lot of work on DNA-peptide concatamers
- Took a flat/bland peptide backbone, added amino acids; trying to get reproducible peptide signal
- Even a not very good sequence
- In some ways the bar is lower than DNA sequencing
- More efficient to look at concatamers
- Starting to generate unique profiles, very reproducible, remarkable how tight the pair profiles are
- Where does it lead to?
- The value is in the hard bits
- Trypsin digest
- Make a conjugate
- Sequence
- Method can deal with partial digest - long peptides and overlapping peptides
- very difficult, don't want to over-sell
- Want any lab biologist to be able to run protein sequencing on the same hardware, no special box needed
- Have a good concept for unique reproducible ...
- We think we're on the first rung of the ladder
- Outy based, can re-read peptide, can pull out of the pore, drop it back, pull it out again, then spit it out when done with it
- Summary
- Trying to develop methods to look at single molecules from blood / water
Q&A
-
How does R10 perform on direct RNA?
- Very enthusiastic, but small following, but scientifically a growth area
- About to embark on refactoring to be compatible with 10.4; may become outy-based method
- Expect a new version, may incorporate re-reading
-
Does outy sequencing method have impact on read length?
- Primarily focused on 3-5kb right now
-
Duplex reads are awesome, but efficiency is low; what's up?
- Seeing now upgrades to duplex, will get higher and higher yields
- 5% -> 20% -> 40% -> 70%
-
Peptide sequencing, what are the longest, does DNA motor slip?
- 8-12 peptides, can increase using a longer pore
- Some motors can move over the peptide, still active development
- Longer is better
Welcome Back To London Calling 2022
Zoe McDougall, Oxford Nanopore Technologies
- Final day of London calling talks
- 5,000 people registered to attend online
- Track from Bowie's best album, changes
- Conversation from Sarah (10x genomics)
- Growing, but want to retain foundational, deep values
- Community has developed new tools, provided feedback to improve whole user experience
- Anything, anyone, anywhere means enabling more scientists
- Goal is to grow without changing anything that we're doing in the community
- Remarkable work to drive single molecule raw read accuracy 99.6% simplex, 99.92% duplex
- What could you do with 72kb perfect reads?
- 6,000 Trillion bases sequenced on nanopore devices [in last year?]
- Surge of sequencing and publications from COVID
- Meaningful increase in cancer and metagenomics
- More than 3,000 publications; grass-roots movement
- Manufacturing building is built in the shape of a MinION
- New manufacturing centres 2022
- Particularly proud of new dots in Africa and South America
- How do we grow?
- We think we need to listen; everyone wants to hear from you
- Ask any nanopore person, they will find the right person to speak to you
- Platform that is so versatile that it can be used for all your needs
Understanding cancer epigenetics, immunogenetics, and energetics
Michael Dean, National Cancer Institute, USA
[Pink Floyd - Another brick in the wall]
- I despise cancer; took the life of relatives; including wife's mother - the grandmother of my children
- Most of career on an army base - resources given to war on cancer
- War is the wrong analogy
- Current analogy - cancer moonshot
- We knew where the moon was; it was a known problem
- Cancer is not a known problem; it's a puzzle that is needed to be solved before finding a solution
- Hundreds of puzzles, each need to be solved
- Since 1990, mortality of breast cancer cut in half, mostly from improved therapy
- Figured out nearly all causes of cancer
- Chimney sweeps were exposed to chemicals
- Genes
- What do they all have in common?
- NBCCS - part of international team that cloned a gene; chr9; human homolog involved in embryonic development (abnormal wings)
- Found many genes critical for embryonic development
- Any time a cell divides, it makes about 100 mistakes; stem cells accumulate mistakes
- But that was the easy part
- Lab focused on cervical cancer; caused almost exclusively by HPV
- Mortality climbs exponentially as income goes down
- Studied in guatemala; tumour and blood samples collected
- HPV is ds virus; often integrates into genome; have used MinION to study cancer
- Last year, showed ultra-long concatamers of HPV sequence
- Order arranged differently at different sites - HPV superspreading
- Found category of HPV that was only extrasomal
- Found other integrations with rearranged multimers
- NCI did in-situ hybridisation; largest fragments of HPV ever seen
- "Nothing in life is to be feared, everything is to be understood" - Marie Curie
- Virtually all cancer patients die of metastatic disease, or therapy-resistant tumours
- New class of genes mutated in cancer
- Modify histones, regulate splicing, regulate methylation
- Theory is that cells with same DNA change into different cell types by regulating the epigenome
- Parallel that onto cancer cell - ability to unlock all pathways of survival
- Angiogenesis
- All our cells are resistant to our own immune system; cancer cells are doing the same thing
- Did direct cDNA sequencing on MinION
- Unprecedented detail on HPV
- Only 1 of 6/800 copies was expressed, makes up majority of transcripts making E7 oncogene
- Other of 1000 has 2 copies making oncogenes
- HPV can integrate inside genes (e.g. FOXE1)
- Gene highly expressed near integration site
- Can see that FOXE1 is overexpressed, mostly not full-length transcripts; alternatively spliced, region downstream that is normally not expressed
- Looked at how integration is affected by methylation
- Can see every CpG site along virus
- Caski cells have more methylation
- FoxE1 has demethylation
- Can see every CpG site along virus
- Direct RNA sequencing
- RNA used to make DNA copy, just to stabilise molecule
- Can determine size of polyA tail
- Applied to HeLa cells (has NYT bestseller)
- Soft-clipped fragments aligning to chr8, MYC oncogene overexpressed
- Would like to translate knowledge into the clinic
- PI3K signaling pathway is frequently affected
- PD-L1 inhibited by PI3K drug, turning off immunosuppressive signals, as well as reducing proliferation
- Treated cells have RNA expression of HPV reduced
- Epigenome is not a linear thing, 3D interactions that we need to understand
- Can use Pore-C to study
- Can see that most of HPV-human interactions are happening in enhancer region
- Cancer cell energetics is dramatically altered
- Warburg - switch from glucose metabolism to lactate
- Full-length reads across whole mt genome
- In general little methylation of mtDNA
- Can count up transcripts from direct RNA
- See Rossi et al. BioRxiv 2021.10.22.465367v1
- Can we develop an inexpensive HPV point of care device?
- Some conversations already
Q&A by Lauren Sunda
-
What are the pros and cons of cDNA vs direct RNA?
- Results from both are highly comparable
- More output from cDNA
- Still exploring RNA modifications, don't know what we're going to find
-
A lot of countries don't motivate women to get test before age 30; how do we deal with that information, it might create more fear?
- Education is the key
- A self-sampling model using a swab; hopefully will be able to apply that better
- Women need to get daughters vaccinated
-
What is the mechanism of hypomethylation around integration sites?
- Is demethylation before or after integration?
- Possibly enhancer brought in by HPV that leads to demethylation
-
Stage experience - Wow.
-
Third trip to London
-
Be bold, curious to follow your instincts
-
Working in Guatemala was challenging
Spotlight Session; runners up will be presenting talks in mini theatre
- STEM students joining this session; part of full-day STEM event to inspire next generation of scientists
Whole mitogenome variation in domestic cats discovered via nanopore sequencing [video]
Emily Patterson, University of Leicester
- Cat hair can be used as a forensic tool; used in July 2012, found wrapped in shower curtain; 8 cat hairs found, police wanted to know if they could be linked to the suspect
- Domestic cats are a very popular household pet
- Frequently found at crime scenes, but limited visual benefit
- Have to depend on mitochondrial DNA - higher copy number, greater stability
- Traditionally sequence 402bp sequence in control region, has repetitive sequences either size
- Cats have quite a large NmMT
- Once sequenced, assess profile in a population database
- 12 universal mitotypes worldwide, 60% cats have 1 of 4 common mitotypes
- Cat population database from 102 cats at Uni of Leicester
- Cat from shower curtain matched rare profile, matched suspect's cats
- Evidence of increased discrimination by sequencing different mt genome regions
- Wanted to build a new database from 93 cat blood samples, 2x overlapping amplicons avoiding NuMT
- Also sequenced using Illumina sequencing, used variation identified to design primers for whole mitogenome
- Used native barcoding, sequenced 8 cats per flow cell
- Good agreement between Illumina and Nanopore; 10 INDELs that were excluded
- 422 polymorphic sites outside repetitive regions
- Aligned cats to sanger Sequence reference sequence, quite a long branch with 42 sites unique to reference
- 3x unique profiles from control region, now have 60 unique mitotypes, reduced largest group from 20 cats down to 4
- Wanted to apply this to cat hairs, needed shorter amplicons
- Average length 360bp amplicons (60 different amplicons); sequenced matched blood and hair samples
- Successfully recovered complete mitogenome from blood and hair
- Variants called in blood have NuMT prominent, found in several cat species
- Applied to missing cat case; tried to match jawbone to missing cat
- Complete mitogenome sequence, found a match
- Allowed to distinguish 74 haplotypes, reduced frequency of largest haplotype from 26% to 4%
Q&A
-
Hardly any variation in 12S gene; any relevance?
- domestic; quite recent; don't know why 12S
-
How did you isolate DNA from cat hairs?
- Qiagen mini kit; used Qiagen protocol for blood
-
How many cat hairs do you need to isolate DNA?
- We used around 10
-
Stage experience
- Quite intimidating; everyone is being really helpful
-
What motivates you?
- Very direct application, can lose sight of things
-
No cats at home, but no parents
-
Obstacle in research - COVID, out of lab for 6 months, had to adapt project; started out as a side project
Animal conservation in the Era of genomics
Genomic tools for gorilla population dynamics and conservation
Etore Fedele, University of Leicester, UK
- We are living through one of the worst extinction events in the
history of the planet; gorillas are no exception
- Poaching, habitat loss, diseases, political issues, bushmeat
- Need to come up with an effective way to monitor changes, identify subspecies
- Want to estimate population size
- Advent of HTS have allowed access to whole range of genetic markers
for identitification
- Single base changes, require only short fragments to be analysed
- Want to reduce risk of spreading infectious diseases
- Issue: very high startup fees for laboratory equipment; samples have to be shipped from Africa to sequencing sites; international regulation to limit transport of biological samples
- MiniPCR + bento allow library preparation in the field
- MinION enables researchers to conduct sequencing in remote areas
- Workflow
- Utilised public data for constant SNPs + heterozygous SNPs
- Designed primers
- Assessed performance by matched blood and faecal samples
- Used both MinION and Flongle side-by-size
- Guppy -> Minimap2 -> BCFTools Results from constant population SNPs + 28 individual SNPs
- Fragment length - really short bases
- Three patterns - western gorillas + 2 others
- Individual identification SNPs, each row different from the others
Flongle sequencing reveals cryptic amphibian diversity and builds local research capacity in the Ecuadorian Amazon
Zane Libke
- 8 years old when caught first snake, filled with fascination about secrets, this fascination led me into the jungles of Ecuador
- Originally tried looking at snake diversity, changed to frogs, found heaps of biodiversity
- Current counts 140 species of mammals, 105 reptiles, 700 birds, 149+ species amphibians, many more
- Area is not a typical Amazon area; incredibly mountainous
- Amazing network of researchers and conservation activists
- Read paper returning from first Ecuador trip
- 1 table has all equipment we use - MiniPCR + MinION
- Lab is product of the pandemic - what better to do than construct a lab
- 1 solar panel powers all research; lots of didactic material
- Protocol adapted and simplified as much as possible, making the best out of every penny
- Extraction for 7 cents
- Barcoding from as many species as possible
- Expedition to 2 different mountains
- River separating, found 2 different but similar frogs; both didn't have a BLAST hit (likely represent undiscovered species); didn't look like each other
- Information passed onto collaborators
- This kind of species resolution can inform conservation decisions
- Diversity is unprecedented
- Out of 44 different morphospecies, found 18 candidate species
- How could we know how to best conserve them?
- Nanopore is a diverse device, allows many things
- Have so far only looked at species we know
- A need to create more labs like we're doing
- Started conducting field genetics courses; 3-7 day intensive course; need more students; huge demand, need for this research
- A lot of ideas and collaboration come out of the courses
- Things learned, will be touched on during plenary talk
Genomics in the jungle: a field laboratory success story
Mrinalini Erkenswick Watsa, San Diego Zoo Wildlife Alliance, USA
- Can't be with us in London, pre-recorded talk
- Focus on field genomics - making it practical, affordable
- Living planet index - last 50 years 68% decline, higher in some areas
- Want to say if conservation monitoring and program is successful
- Monitoring is spread in a biased fashion across the world; samples get lost, degrade; exporting samples exports opportunity
- 2020 - consortium of researchers advocated for decentralised labs; nanopore technology's greatest advantage
- The In Situ Laboratory Initiative
- First node in south-eastern Peru (team with decade of experience)
- Other hubs emerging globally, Rwanda, Kenya
- Los amigos; enhanced solar capabilities, codified spaces, OpenTrons
- BSL-2 lab in the middle of the Amazon rainforest
- Case 1 - Every sample being DNA barcoded; have multiplexed 11 plates onto one run using dual barcoding
- Case 2 - Population monitoring; optimised for sample collection
- Case 3 - Viral / disease surveillance; have identified one novel coronavirus strain
Endangered European sturgeon detection through non-amplified eDNA sequencing
Reindert Nijland
- Getting full genome of European sturgeon
- The Netherlands have a number of species of sturgeon, most escaped from ponds and aquaria
- We have sturgeons, but not the species we'd like to see
- Pretty enigmatic, use both freshwater and seawater, would be nice to see these really big fish back in European waters
- Population living in France have been collected, hope to reintroduce before it goes completely
- Developing method; started with 3 swimming animals as part of reintroduction program
- Support from nanopore via org.one; isolated DNA with NEB Monarch,
LSK110 + ULK kit
- Draft genome sequence is helping us a lot
- Have 1.8 Gb of sequence (whole genome), would previously only have marker genes
- No more PCR needed to detect environmental DNA
- Teaching course; masters students decided to sample sturgeon tank
- PCR on marker of sturgeon, found target
- Collected DNA, 1.5ng in total volume
- Thought maybe kit 12 could do something
- It did work, throughput was very low, previously used flow cell; a couple of really nice reads in 1/2 hour, matched target and could tell apart from another sturgeon
- Re-did experiemnt, about 1% of environmental genes matched sturgeon
- Currently trying to sequence genome of other species
- Great proof of concept, hope to extend to other species of fish, then maybe extend to biodiversity as a whole
- Historical records showed 2x sturgeon alive in The Netherlands; could be that all the historical specimens have not been identified correctly; new genomes will help
Q&A - Mrinalini joining from Zoom
-
Museum specimens - can you sample the fluid?
- Could possibly do it from the ethanol, but that has been replaced
-
Extraction techniques, cost, sample limitations
- Mrinalini - Too much sample to work with; use Open-source pipetting robot; have been using kits that use magnetic beads; make own beads
- Ettore - wanted a kit that could be stored at RT; large used Qiagen kit + chelex - depends on quality of sample; faecal used Qiagen stool kit; Fire tissue kit
- Zane - Don't need to purify DNA for simple PCR (thanks Mrinalini)
- HotShot protocol, can be done in MiniPCR, works for most samples, at least for amphibian and reptiles
- Reindert - Students collected 2L water; Reindert collected 5L; passed over a filter, used Qiagen kit. Don't like PCR for environmental DNA, likes to work with unamplified DNA. North sea try to filter 1.5-2L water (clogs filter); could filter 10L in zoo.
-
Have you used adaptive sampling for eDNA?
- Haven't tried it yet; not sure if it'll work with low concentration
- Mrinalini, Ettore, Zane hasn't tried yet
- Zane has project trying to detect possibly extinct species of toad; don't know where frog exists
-
Adaptive sampling for eDNA for Kakapo - 0.01%, adaptive sampling didn't help; might work with 1%
-
Challenges - in-situ lab; what does it take, how do those decisions get made?
- Mrinalini - biological research station, variety of countries, lots of interesting ideas, incredibly supportive, very interested; making a blueprint for how to do this; it needs community buy-in, opens it up to people who could never do those projects, breaks down so many barriers; we need people to care about it; insects, mould, fungus are a problem (even in boxes of samples)
- Zane - community involvement is key; courses given at research station; still a huge issue acquiring equipment, reagents; one distributor for all of South America
-
Acquisition process for Nanopore stuff
- Zane - first funding was $5k grant from university, bought device, brought on carry-on luggage. Funds can only be used to purchase equiment within the country (Equador), so can't get Nanopore stuff.
- Mrinalini - have been walking around the issue
Spotlight - The bacterial black market: utilising sequencing to study the inter-cellular and intra-cellular transfer of AMR genes in bacteria [video]
Richard Goodman, Liverpool School of Tropical Medicine, UK
- Less on the Star Wars
- AMR is a huge threat globally, UK government in 2016 reported deaths
by 2050 would reach 50 million (if no intervention whatsoever)
- 2019 alone, 4.9 million deaths from AMR, knocked estimates out of the water; carried mostly by low and middle-income countries
- How do bacteria acquire AMR? Point mutations, HGT - share pieces of their DNA between themselves - transformation, transduction, conjugation
- Humans have millions of bacteria on each surface, some contain antimicrobial resistance genes; picture that the environment is more complex than described
- Intercellular transfer - DNA moves between cells; intracellular transfer - DNA moves off plasmids onto genome; getting gene onto genome is more energy efficient
- Bacteria gets a resistance gene, give antibiotics, leaving one with AMR gene
- patients in hospital are pushed towards a monoculture with lots of resistance genes
- Colistin: the last resort [after carbapenems]; a polypeptide that binds to LPS and solubilises cells
- First resistance was point mutation
- Then found AMR gene on plasmid; adds phosphoethanolamine to lipid A
- Looking at this movement of genetic elements
- Typical approach - agar plate; looking at bacterial genome
- Developed entrapment vector that captures mobile elements
- Antibiotic repressor gene; tetracycline gene
- Sequencing to look at what's happening
- Initially looked at intracellular transfer; transposon with mcr-1 gene
- Looked at different resistances
- Were able to check resistance on a plate
- Need sequencing because plating doesn't say where resistance is happening
- Nanopore sequencing is rapid and accurate, can distinguish plasmid from genomic
- Helps work out how they jumped into the chromosome
- Assembled with Flye, polished
- Found backpack plasmid jumped into the chromosome, mediated by ISApl1
- Entrapment vector was captured itself; now defined as a transposon
- Were also able to show the movement of this transposon, completely novel movement
- Detected lots of different insertions
- Wanting to see if this can translate into therapy
Q&A
- Species - mainly E. coli
- Research could be used in the field
Unlocking the transcriptomic architecture of bacterial viruses with ONT-cappable-seq [video]
Leena Putzeys, KU Leuven, Belgium
- Phages are viruses that specifically infect and kill bacteria; a valuable source of inspiration for biotechnical applications
- They do it so efficiently; regulate gene expression; start transcription in a very systematic way
- Subset of early genes to take over host
- Replicate phage
- Viral assembly & lysis
- Trying to understand mechanisms
- Phage transcriptomics is mainly driven by Illumina, very compact genomes, very dense, difficult to analyse on a transcriptomic level
- Illumina coverage is very variable, very hard to make out what is happening; what is the original composition
- Classic RNA sequences do not discriminate between primary and processed transcripts
- Interested in primary transcripts, with original bindings, help in finding original transcription sites
- Cappable-seq methods; primary transcripts have triphosphate at end; can use 5'G cap
- allows to sequence full length
- Calling approach ONT-cappable-seq
- Different stages of infection
- Implemented in-vitro polyadenylation
- One sample enriched, other as control
- Used second-strand cDNA synthesis
- Compared RNA profiles from Bioanalyser before and after
- high peaks, high amounts of processed RNA; enrichment affects this; loses peaks
- Control sample - most mapping to rRNA
- Interested in finding out transcriptional boundaries
- Transcriptome for ONT-cappable-seq makes it much easier to see where transcripts start and end; initiation site, termination site
- Interesting patterns
- Region with transcriptional readthrough with multiple terminators (can incorporate up to 5 additional genes)
- Intergenic region with a lot of transcriptional activity, small RNA species
Q&A
-
Approaching this for the first time?
- Lots of tweaking needed; give it a go and see
- Make sure you have nicely intact RNA
-
What kit used with PromethION?
- PCB109; 12 barcodes with R9 chemistry
Post-conference review - Long-read transcriptome sequencing reveals isoform diversity across human neurodevelopment and aging [video]
Rosemary Bamford, University of Exeter, UK
Post-conference review - Applying nanopore sequencing in the undergraduate classroom for environmental microbiome analysis [video]
Ashley Beck, Carroll College, USA
- Educator & scientist - research & tech to next generation
- Nanopore past couple of years seq into hands of undergraduate students
- Study sites - microbial communities; took samples from contaminated river
- Collected from locations along river, extracted DNA from sediments
- Next years, students designed own projects
- Drinking water (good; filter should be changed)
- Soil microbial commmunities around smelter site
- Streamlined workflow; 16S sequencing kit to multiplex on MinION Mk1B
- Examined samples via Epi2Me
- Challenge of exploring data later on in R later
- Scientific results - could get all way through project; distinct differences for smelter sites
- Got Preparation in being able to see project from design to presentation (1 presented at NCM)
- Survey assessment for students - creativity; discovery & relevance; science
- scores relatively high
- scores went up for second year
- "Exposure to new techniques and technologies helped me to feel less nervous and lost"
Post-conference review - CyclomicsSeq: targeted and genome-wide detection of circulating tumour DNA using nanopore consensus sequencing [video]
Jeroen de Ridder, Cyclomics & University Medical Center Utrecht, Netherlands
Post-conference review - Novel method for multiplexed, full-length single-molecule sequencing of human mitochondrial DNA using Cas9-mediated enrichment [video]
Ivo Gut, CNAG-CRG, Spain
Post-conference review - Nanopore sequencing shows potential for personalised oncogenomics [video]
Kieran O’Neill, Canada's Michael Smith Genome Sciences Centre / BC Cancer, Canada
Post-conference review - De novo assembly of immunoglobulin loci linked to full-length single-cell transcriptome of antigen-specific plasmablasts [video]
Christian Stevens, Icahn School of Medicine at Mount Sinai, USA
-
Subjects with MMR vaccine boost
-
Day 6 - isolate PBMCs - B-cells single cell; non B-cells get germline DNA for IG loci
-
Single cell transcriptomics with ONT only (sockeye)
-
B-cell Development
- Antibodies are built to stick to things
- Variable region (sticks)
- Constant - effector functions
- Human DNA only have so many bases to work with, human capacity up to 1M - 1T different combinations
- Need to recognise any non-self antigen that could exist
-
MMR vaccine
- Trivalent, 3 inactivated viruses
-
Antibody
- Heavy chain, light chain
- All cells have germline loci
- One B-cell makes only one antibody
- mu-VDJ recombination (heavy), then light-chain (kappa or lambda) recombination
-
A lot of V/D/J genes; Single B-cell selects one of each of those
-
RNA splicing gives constant region
-
Antigen comes in, B-cell waiting in the wings; isotype switching and somatic hypermutation
-
Antibody effects include downstream function; e.g. neutralisation, opsonisation
-
Lastly somatic hypermutation; each new antigen selects higher and higher affinity to antibodies
-
Sorting
- Different surface expression of cell proteins, flow sorting
- Differentiate plasmablast and B cells, looking for memory B cells
-
10X sequencing for transcriptomics
-
ONT genomic applications team - extracted DNA with IG loci adaptive sampling (50Mb targets) + immune genes, about 60X coverage of IG loci; assembling germline allows even more that can be done
-
Can phase and get haplotypes
-
Single phase-block haplotypes
-
Assembly
- Transcript from 10X -> find V/D/J
- Found potentially novel new genes
- novel V genes creates qualitatively different loci
- Found 4 individual alleles associated with memory B cells
- Subset of cells with heavy chain B allele as a novel locus
-
Comparing Illumina vs Nanopore
- ONT vs Illumina expression levels (7 cells), very homogenous cell population, cell barcode agreed between ONT and Illumina
- Illumina and ONT are capturing the same thing
- Illumina only, looking only at IGHG genes; IGHG1 vs IGHG2, no correlation
- For IGH1/3/4, there is some overlap
-
Why not Illumina?
- Nearly identical isotypes, except at hinge region
- hinge region is a way down from the read, so Illumina is not successfully picking up the difference
-
ONT
- Not seeing the same correlation for IGHG isotypes
- ONT vs Illumina has poor correlation; demonstrates mis-assignment for Illumina
- Can cluster plasmablasts based on IGH isotypes
-
Why does seeing differences matter?
- B4GALT1 - enzyme involved in glycosylation of antibodies; antibody that can affect ATCC activity
- Didn’t expect to see different glycosylation factors as important
- Significant enrichment in IGHG1 isotype
- Can do differential expression analysis, comparing IGHG subsets
-
Not Just any old gene, but does happen with others
- B4GALT1 is the only 1 showing up
- B4GALT1 is showing different effector functions
- Being able to pull that out is an exciting part of this project
-
Additional:
- Pulled antibodies from specific plasmablasts
- Synthesis in China; lockdown was a problem
- 3 synthesised from potentially novel allele
-
Can get not just VDJ recombination, but exact isotype
Post-conference review - HLA typing using targeted third-generation sequencing methods [video]
Steven Verbruggen, OHMX.bio, Belgium
Post-conference review - Portable real-time sequencing to safeguard critically endangered wildlife [video]
[walk-on music - NZ bush, with bird calls]
Marissa
- NZ should sound like a land filled with diverse bird life
- Doesn't sound like that, due to mammalian predators, absolutely devastating birdlife; birds evolved without mammalian predators
- Kakapo has suffered severely
- Flightless, nocturnal
- Population down to 51 individuals in 90s; now 197 (down 2 from when slides were made)
- Intensive management working really well
- Slow-breeding species, every 2-4 years
- Intensive management has made this a good system, could be applied to other species
- Every kakapo chick that hatches is precious
- Weigh chicks every 2 days
- Male / female have slightly different weight gains; need to tell female vs male in order to work out if weight gain is normal
- Kakapo don't have a lot of genetic diversity left
- Artificial insemination used to increase representation
- Many founders from 90s are still alive today, trying to make sure they still have offspring
- Aim: develop and deploy a rapid test for parentage and sex
- Doing with portable nanopore sequencing; portability was essential; affordability is also useful
- Want protocol that can be left with the team after leaving
- Protocol shown to rangers; they get excited about the technology
- As few cold chain requirements as possible
- For low coverage sequencing, how much data is needed?
- Downsampled data to various yields
- Only need about 1.6 Gb to get good parentage information
- Any yield tested generated useful results for sex determination
- Criteria: wanted to make sure regardless of samples, would be robust and give good results
- Went out to nests at night time
- Collected blood sample from chicks after mum had left (capillary)
- Qiagen spin column-based kit
- Bento lab
- Insulated plastic box + kettle water
- Rapid barcoding kit, MinION flow cell
- Magnetic rack
- Nvidia portable GPU
- MinION
- Setup was great
- Could read off within an hour of sequencing
- Where used?
- Pearl A2; initially doing well, but started to tail off
- a couple of days later, found out chick was male; more concerning
- Brought chick back from nest, was able to make a recovery with vetinary care
- Sex results for all kakapo chicks
- Parentage results for important birds; showed that artificial insemination was successful
- Some chicks descendants of a founder that was getting very old
Lara Urban
- Used in-situ & real-time component of nanopore sequencing to conduct portable wildlife epidemiology
- Picture: kakapo chick cuddling MinION sequencer
- Critically-endangered species is already at the brink of extinction
- Can try to support species and protect from disease, but need to understand disease risk
- Outbreak of aspergillosis last year; normally occurs everywhere in the environment; killed a few chicks (every one is important)
- Needed a fast in-situ monitoring of this strain, wanted to give techniques to ranges
- Needed to establish protocol using nanopore metagenomics from laryngal swabs
- used Copan eSwabs
- extracted DNA; needed to break fungal cell walls with handheld pestler & glass beads
- found aspergillis, potentially quantitative
- Used rapid barcoding kit, useful due to transposase
- Sequenced libraries, 12 samples for around 12 hours; using Jetson Nano
- That's the theory
- Kakapo mum staring at screen, waiting for kakapo mum to leave, would get excited; need to wait for motion sensor to make sure mum had left (loud ring signal)
- Get chick out of nest, act like feeding chick until they reacted
- Miles Benton has made powerful embedded system work with Nanopore sequencing
- Jetson board AGS has powerful GPUs
- Did live alignment of sequencing reads
- Was able to try adaptive sampling in the field for depleting host
- Results - understand baseline; did a timeline of two different chicks (Pearl-A1, Pearl-A2)
- There is a baseline load of aspergillis (< 0.5% of reads)
- Drop between day 20 for Pearl-A2; mother decided to focus energy on other chick
- Base load probably coming from mother's food
- Sequenced quite a few more samples, so far haven't detected any aspergillis reads in chicks
- Don't hope that this will happen, but needed to validate method
- Tried adaptive sampling, number of reads slighty reduced, number of base pairs about half; Aspergillis load approximately doubled
- Host depletion might allow detection rate to be increased
- This was a pilot study; generally very interested in using nanopore sequencing to improve planetary health
- Use for wildlife conservation
- Have previously used nanopore sequencing for monitoring environment
- Want to bring this all together, assess impact on the environment
- Currently setting up own research group at Helmholtz AI
- Thank Kakapo Aspergillosis Research Corporation; ONT (sent Mk1Bs + more consumables)