Arichard:Notebook/AR151201: Difference between revisions
>Andrew (Created page with "==Alginate GMD encapsulated nuclei on DropSeq generator== December 1, 2015 Experiment code: AR151201") |
>Andrew |
||
(6 intermediate revisions by the same user not shown) | |||
Line 4: | Line 4: | ||
Experiment code: AR151201 | Experiment code: AR151201 | ||
===Motivation=== | |||
Contiguity Preserving Tagmentation (CPT-seq) was demonstrated by Illumina in 2014 for haplotyping: | |||
http://www.ncbi.nlm.nih.gov/pubmed/25326703 | |||
The enabling concept is that tagmentation of double stranded DNA takes place in multiple steps. Tn5 binds rapidly to double stranded DNA, but gel electrophoresis shows that fragments do not appear until after Tn5 is removed from the template by SDS, EDTA, or heat. The intermediate step, between binding and fragmentation, resembles a bead-on-a-string configuration. It is therefore possible to transfer long, double-stranded DNA fragments loaded to saturation with Tn5 transposome complexes from one reaction to another, e.g., from tagmentation to PCR. CPT-seq uses barcoded transposons in a 96-well plate, followed by random mixing and re-distribution into a second 96-well plate with barcoded PCR primers, followed by size-selection and sequencing. This creates a very large number barcode space very few oligos. A low desired collision rate (e.g., 0.1%) is achieved by randomly under-loading the total barcode space (e.g., 1 - 10%). Fragments with identical barcode combinations can be inferred to have originated from the same long fragment. | |||
Combinatorial Tn5 barcoding was adapted for chromatin accessibility by the Shendure group (combinatorial scATAC-seq) by starting with single-nuclei suspension instead of DNA in solution. Extracted nuclei are sufficiently permeable to Tn5, such that accessible chromatin can be tagged with barcoded transposons. The nuclei can then be pooled and split for barcoded PCR. | |||
The strength of each of these methods lies in the selective and reversible co-transfer of material from one step to the next. CPT-seq co-transfers densely-tagged, long single fragments from many genomes, while combinatorial scATAC-seq co-transfers sparsely-tagged, whole single-cell genomes comprised of many long fragments with native chromatin structure, i.e., chromosomes. Adapting a combinatorial Tn5 approach to generate high-coverage, whole-genome, single-cell libraries for CNV calling requires the co-transfer of fully accessible single-cell genomes, yet most methods of accessibilizing genomic DNA destroy the cellular and/or nuclear structure that combinatorial scATAC-seq relies on for co-transfer. | |||
This experiment is based on two published methods, Drop-Seq and surfactant-free monodisperse alginate GMD generation: | |||
http://www.cell.com/abstract/S0092-8674(15)00549-8 | |||
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3505195/ | |||
First try had very few, very large droplets, incompletely polymerized: | |||
[[File:AR151201_FirstTryTooFast.jpg|600px]] | |||
The GMD paper indicates that there are three domains of flow: streaming, plug, and droplet: | |||
[[File:flowDomains.png|600px]] | |||
Based on the following figure, we decided to reduce the aqueous phase flow-rate: | |||
[[File:AR151201_Plugs.png|600px]] | |||
We then saw plugs, which indicate that the aqueous flow rate is still too high compared to the organic phase, but that there is sufficient time for polymerization, assuming that an unpolymerized solution would form a sphere in an organic phase after leaving the device. | |||
We then reduced the flow rate to 500 ul/hr: | |||
[[File:AR151201_GettingRounder500aq.png|600px]] | |||
Getting better. | |||
Reduced flow rate to 100 ul/hr: | |||
[[File:AR151201_Drops.png|600px]] | |||
Now we have droplets, estimated to be 60 - 70 um in diameter. | |||
We poked them with pulled glass capillaries to confirm that they were gelled. | |||
[[File:AR151201_Poking_Drops.png|600px]] |
Latest revision as of 18:39, 11 December 2015
Alginate GMD encapsulated nuclei on DropSeq generator[edit]
December 1, 2015
Experiment code: AR151201
Motivation[edit]
Contiguity Preserving Tagmentation (CPT-seq) was demonstrated by Illumina in 2014 for haplotyping:
http://www.ncbi.nlm.nih.gov/pubmed/25326703
The enabling concept is that tagmentation of double stranded DNA takes place in multiple steps. Tn5 binds rapidly to double stranded DNA, but gel electrophoresis shows that fragments do not appear until after Tn5 is removed from the template by SDS, EDTA, or heat. The intermediate step, between binding and fragmentation, resembles a bead-on-a-string configuration. It is therefore possible to transfer long, double-stranded DNA fragments loaded to saturation with Tn5 transposome complexes from one reaction to another, e.g., from tagmentation to PCR. CPT-seq uses barcoded transposons in a 96-well plate, followed by random mixing and re-distribution into a second 96-well plate with barcoded PCR primers, followed by size-selection and sequencing. This creates a very large number barcode space very few oligos. A low desired collision rate (e.g., 0.1%) is achieved by randomly under-loading the total barcode space (e.g., 1 - 10%). Fragments with identical barcode combinations can be inferred to have originated from the same long fragment.
Combinatorial Tn5 barcoding was adapted for chromatin accessibility by the Shendure group (combinatorial scATAC-seq) by starting with single-nuclei suspension instead of DNA in solution. Extracted nuclei are sufficiently permeable to Tn5, such that accessible chromatin can be tagged with barcoded transposons. The nuclei can then be pooled and split for barcoded PCR.
The strength of each of these methods lies in the selective and reversible co-transfer of material from one step to the next. CPT-seq co-transfers densely-tagged, long single fragments from many genomes, while combinatorial scATAC-seq co-transfers sparsely-tagged, whole single-cell genomes comprised of many long fragments with native chromatin structure, i.e., chromosomes. Adapting a combinatorial Tn5 approach to generate high-coverage, whole-genome, single-cell libraries for CNV calling requires the co-transfer of fully accessible single-cell genomes, yet most methods of accessibilizing genomic DNA destroy the cellular and/or nuclear structure that combinatorial scATAC-seq relies on for co-transfer.
This experiment is based on two published methods, Drop-Seq and surfactant-free monodisperse alginate GMD generation:
http://www.cell.com/abstract/S0092-8674(15)00549-8 http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3505195/
First try had very few, very large droplets, incompletely polymerized:
File:AR151201 FirstTryTooFast.jpg
The GMD paper indicates that there are three domains of flow: streaming, plug, and droplet:
Based on the following figure, we decided to reduce the aqueous phase flow-rate:
We then saw plugs, which indicate that the aqueous flow rate is still too high compared to the organic phase, but that there is sufficient time for polymerization, assuming that an unpolymerized solution would form a sphere in an organic phase after leaving the device.
We then reduced the flow rate to 500 ul/hr:
File:AR151201 GettingRounder500aq.png
Getting better.
Reduced flow rate to 100 ul/hr:
Now we have droplets, estimated to be 60 - 70 um in diameter.
We poked them with pulled glass capillaries to confirm that they were gelled.