Daniel:Notebook/ComboLock/2017-5-16: Difference between revisions

From ZhangLabWiki
Jump to navigation Jump to search
>Djacobse
>Djacobse
 
Line 107: Line 107:
File:Poscon.tf-basecomplexity.png|Positive Control (TF) base complexity
File:Poscon.tf-basecomplexity.png|Positive Control (TF) base complexity
File:RandomUMIs-basecomplexity.png|Randomly generative UMI base complexity
File:RandomUMIs-basecomplexity.png|Randomly generative UMI base complexity
</gallery>
===Knee Plots (5/19/2017)===
Added from analysis performed yesterday (5/18). The knee plots show the UMIs, ordered by number of reads (descending), vs. total fraction of reads. Only the random UMI samples go up to 1 on the y axis, but that's because the others don't have 100% alignment.
<gallery perrow=3 heights=300px widths=300px caption="Knee Plots for MiSeq 20170508-Fraction of UMIs vs Fraction of Reads">
File:Sample.neb-kneeplot.png|Sample (NEB)
File:Sample.tf-kneeplot.png|Sample (TF)
File:Poscon.neb-kneeplot.png|Positive control (NEB)
File:Poscon.tf-kneeplot.png|Positive control (TF)
File:Dephos.tf-kneeplot.png|Dephosphorylated sample (TF)
File:RandomUMI.250K-kneeplot.png|Random UMI
</gallery>
</gallery>


Line 117: Line 130:
*Dephosphorylated had most (72%) empty UMIs
*Dephosphorylated had most (72%) empty UMIs
*ThermoFisher had slightly better UMI complexity
*ThermoFisher had slightly better UMI complexity
*The knee plots show that the dephosphorylated sample had the highest AUC, which is actually bad for this type of analysis




[[Category:ComboLock]] [[Category:20170428]]
[[Category:ComboLock]] [[Category:20170428]]

Latest revision as of 15:33, 19 May 2017

Production Run (Started Friday 4/28; Libary Prep)[edit]

Back to Calendar

Sequencing Results[edit]

These results are from the sequencing run performed 5/8/2017.

Read Statistics[edit]

Sample Condition Reads Aligned Reads Unaligned Pct Aligned Missing UMIs (%) Poisson Lambda Poissonian E[0] (%) UMI Missing Enrichment
1X NEB Sample 254728 21861 92.1% 34.17 3.89 2.0 16.7
5AX NEB Positive Control 119220 10206 92.1% 31.2 1.82 16.2 1.9
1Y TF Sample 197959 23195 89.5% 42.53 3.02 4.9 8.7
5AY TF Positive Control 63074 6955 90.1% 51.37 0.96 38.3 1.3
4AY TF Dephosphorylated 133660 15412 89.7% 72.47 2.03 13.1 5.5
NA Randomly Generated 250,000 UMIs     2.26 3.81 2.21 1.0

UMI Counting[edit]

Base Complexities[edit]

Knee Plots (5/19/2017)[edit]

Added from analysis performed yesterday (5/18). The knee plots show the UMIs, ordered by number of reads (descending), vs. total fraction of reads. Only the random UMI samples go up to 1 on the y axis, but that's because the others don't have 100% alignment.

Discussion[edit]

Sequencing results are good overall, at least good enough to move on to the BSA positive control and possibly cells for mRNA. Couple of notes:

  • Production run positive control yielded correct sequences
  • Samples had additional empty UMIs to theoretical Poissonian zero
  • Dephosphorylated had most (72%) empty UMIs
  • ThermoFisher had slightly better UMI complexity
  • The knee plots show that the dephosphorylated sample had the highest AUC, which is actually bad for this type of analysis