Tide

TIDE: Tracking of Indels by DEcomposition


Upload Data:

Guide sequence: Submit 20nt guide sequence upstream of PAM ('5-'3)

Parameters

All parameters have default settings but can be adjusted by checking the 'advance settings' box.

Alignment window (bp) The sequence segment used to align the control and test sample left boundary
right boundary automatically set at breaksite - 10bp
Decomposition window (bp) The sequence segment used for decomposition.
Default is maximum window possible

Indel size range Maximum size of insertions and deletions modeled in decomposition

P-value threshold Significance cutoff for decomposition


Created by Bas van Steensel lab
Hosted by NKI
TIDE version 2.0.1
TIDE supports Firefox6, Chrome4, Safari6, IE10 or higher

NOTE: we occasionally receive reports that this webtool does not respond. In some instances this appears to be due to firewall problems. This can often be solved by accessing TIDE from a different computer or network. If you continue to experience problems please contact us and let us know the nature of the problem and the exact date and time when you tried to access TIDE. Your feedback is helpful for us to improve this webtool.

Purpose

This is a web tool to rapidly assess genome editing by CRISPR-Cas9 of a target locus determined by a guide RNA (sgRNA). Based on the quantitative sequence trace data from two standard capillary sequencing reactions the TIDE software quantifies the editing efficacy and identifies the predominant types of insertions and deletions (indels) in the DNA of a targeted cell pool.

Citation

Please cite the article in which TIDE was described in your future publications: Brinkman et al, Nucl. Acids Res. (2014). See the article also for a detailed explanation and examples.

sequence trace

Instructions

1. Upload Data:

Enter a 20nt ('5-'3) DNA character string representing the used sgRNA guide sequence immediately upstream of the PAM sequence (PAM not included). Numbers and other invalid (non-IUPAC) DNA characters will be automatically removed. TIDE assumes that a dsDNA break is induced between nucleotides 17 and 18 in this sequence.
Next, upload the chromatogram sequence files of respectively the control sample (e.g. transfected without Cas9 or without the sgRNA) and the test sample (e.g. DNA of pool of cells treated with both Cas9 and the sgRNA).
We advise to sequence a stretch of DNA ~700bp enclosing the designed editing site. The projected break site should be located preferably ~200bp downstream from the sequencing start site. This region upstream of the break site is used to align the sequencing data of the test sample with that of the control sample.

Currently, ABIF (.ab1) and SCF (.scf) files are supported. SCF is an open standard and several tools exist to convert other formats to SCF files.

2. Enter Parameters for Analysis:

The following parameters have default settings but can be adjusted if necessary in the panel to the left by checking the 'advance settings' box.

Alignment window

These settings determine the window in which the control and test sequences are aligned to determine any offset between the two reads. There is usually no need to deviate from the default settings, except when long repetitive sequences are present.

left boundary: By default this is set to 100, because base-calling at the start of a Sanger sequence read is often of poor quality.
right boundary: This is automatically set at break site - 10bp

Decomposition window

These settings determine the sequence segment used for decomposition. The default setting is the largest window possible for the uploaded sequences:

left boundary: max indel size + 5bp downstream of the break site.
right boundary: max indel size + 5bp before the end of the shortest sequence read.

The decomposition window can be adjusted if part of the sequence read is of low quality or contains repetitive sequences.

Indel size range

Set the maximum size of deletions and insertions to be modeled. The default value is 10.

P-value threshold

Significance cutoff. Any value between 0 and 1 is accepted. Default is p<0.001

3. Results:

Once the data are uploaded and parameters are set, submit the data by clicking on the "update view" button and the plots will appear in the two tabs: Decomposition & "+1 insertion". If the settings are incorrect or too stringent, warnings or remarks will be dispayed in the "Decomposition" tab.

Quality measures: Results depend on the quality of the sequence reads. As a rule of thumb, we recommend to aim for an average aberrant sequence signal strength before the breaksite <10% (both control and test sample), and R2>0.9 for the decomposition result. Sequencing of the opposite strand is recommended to confirm the results.

License, terms of use, privacy

This webtool and the associated R code are open source software under GNU General Public License version 3. If you use this software for data analysis in a publication, please cite (Brinkman et al, Nucl. Acids Res. (2014)). Your uploaded data are only used for the duration of the analysis session and are not stored or used for any other purpose.

Code

Contact Bas van Steensel lab if you are interested in the R functions of TIDE.

Contact

This web tool was developed by Eva Brinkman, Tao Chen and Bas van Steensel. William Peters has assisted with setting up the online web tool. For more information and to report bugs, please contact Bas van Steensel

Acknowledgements

R

R Core Team (2013). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. www.R-project.org . R version 3.1.1.

Biostrings

H. Pages, P. Aboyoun, R. Gentleman and S. DebRoy. Biostrings: String objects representing biological sequences, and matching algorithms. R package version 2.32.1.

sangerseqR

J.T. Hill and B. Demarest (2014). sangerseqR: Tools for Sanger Sequencing Data in R. R package version 1.3.1. http://www.bioconductor.org/packages/devel/bioc/html/sangerseqR.html

nnls

K. M. Mullen and I. H. M. van Stokkum. The Lawson-Hanson algorithm for non-negative least squares (NNLS). R package version 1.4.

shiny

RStudio and Inc. (2013). shiny: Web Application Framework for R. R package version 0.11.1. http://CRAN.R-project.org/package=shiny



A plot will be shown here when the valid sequencing files and guide string have been uploaded.

Indel Spectrum

Quality control - Aberrant sequence signal

Remarks



Checklist quality control

Check the following criteria to determine the quality of your data (use the Aberrant sequence signal plot)

  • Is there a considerable divergent signal between control and test sample after the breaksite?
  • Sequences of good quality show:
    • in the control sample (black) a low and equally distributed aberrant sequence signal
    • in the test sample (green) a low signal before the breaksite and a higher signal downstream of the breaksite

  • Is the breaksite at expected location?
    The aberrant sequence signal should increase around the expected cut site (blue dotted line)

  • Does the decomposition window covers a representative sequence?
    For optimal decomposition, the window is set maximal (default), adjust boundaries when the sequence trace is locally of poor quality
If the plot does not meet these criteria, visit the troubleshooting page to see how you can improve your results.

Quantification Indel Frequencies


The prediction of the +1 inserted nucleotide will be shown here when the valid sequencing files and guide string have been uploaded.

Inserted nucleotide probability (%)

FAQ

Troubleshooting

FAQ

What is the minimal sequence length you need?

The requirements of sequence length are flexible. This region upstream of the break site is used to align the sequencing data of the test sample with that of the control sample. The region behind the break site is used for a decomposition to determine the various indels in the pool of cells. In general, with a bigger stretch of sequence trace a better estimation can be performed by TIDE. We advise to sequence a stretch of DNA ~700bp enclosing the designed editing site. The projected break site should be located preferably ~200bp downstream from the sequencing start site.

What happens if the sequence is shorter than 700 bp?

TIDE should work with shorter sequences if the quality of the sequence reads is good. In that case the start of sequence read (alignment window) might have to be set lower than the default setting of 100. Often with shorter sequences the break site is too close to the start of the sequence read in the default setting (see figure). The alignment window can be changed under Advanced settings. short sequence

Does the break site really need to be 200 bp away from the primers?

TIDE should work with sequences with a break site closer or further than 200 bp to the start of the sequence trace if the quality of the sequence reads is good. In that case of break site closer to beginning, the start of sequence read (alignment window) have to be set lower than the default setting of 100. The break site is too close to the start of the sequence read in the default setting to perform an alignment (see figure). The alignment window can be changed under Advanced settings. In case of break site further from the beginning, the decomposition window become smaller. Note, a larger stretch of nucleotides selected often makes the estimation of the indels more reliable. The minimal decomposition window spans from 2x n bp indel size & it has to be at least 5bp from the sgRNA target site or the end of the sequence. short sequence

How is the overall efficiency calculated?

The overall efficiency doesn't need to add up to 100%, because there is also noise in the data.
The overall efficiency is calculated as R2 - % wildtype (zero indels). For example if the R2 value is 0.95, it tells that 95% of the variance can be explained by the model; the remainder 5% is noise or very large indels.

What do the different indel bars indicate when cell pool was sequenced?

The different indel bars represent the different insertion and deletions in the population. You can’t tell for an individual cell what the specific indels of each allel is. To determine allel specific information you have to isolate a cell clone and perform TIDE analysis.

What do the different indel bars indicate when cell clone was sequenced?

The different indel bars represent the different insertion and deletions in the allels in a cell clone. With a diploid cell you should get a percentage of ~50% per indel.

Can we get the sequence of different indels?

The sequences of the specific indels cannot be deduced from the sequence trace. We tested this option, but Sanger sequening is not sensitive enough to give reliable results on this. To know the precise sequence of the mutations you can use next generation sequencing or sequencing the DNA of individual mutated clones.

Can TIDE be used for other nucleases (e.g. TALEN, ZFN, other RNA guided nucleases with different PAM)?

You can use TIDE for other RNA guided nucleases by entering in the TIDE webtool the DNA sequence around the expected cut site. TIDE assumes that a dsDNA break is induced between nucleotides 17 and 18 in the input sequence. You can estimate which sequence this will be for the other nucleases. Note that if you don’t know the exact breakpoint, TIDE should work fine to estimate the efficiency, but the +1 estimation is not be reliable anymore.

Troubleshooting

TIDE web page turns grey

In some instances the webtool appears not to respond due to firewall problems. This can often be solved by accessing TIDE from a different computer or network. If you continue to experience problems please contact us and let us know the nature of the problem and the exact date and time when you tried to access TIDE. Your feedback is helpful for us to improve this webtool.

Webtool does not respond when uploading .ab1 files

At the moment some .ab1 sequence files are not compatible with the TIDE webtool. Various software programs that process the raw sequence data can cause this problem. We are working on this issue. For now, we recommend that you export the data as a .scf file and then uploaded in TIDE. The .ab1 format can also be converted to .scf using software such as 4peaks (Mac) or FinchTV (Windows & Mac).

Wrongly annotated nucleotides

Sometimes the quality of the peaks in chromatogram looks fine, but the file has some wrongly unannotated or wrongly additional annotated nucleotides. These will interfere with the indel spectra (see figure wrongly unannotated nucleotide). TIDE gives a warning when the spacing between the nucleotides in the chromatogram of the sequence trace are not consistent, which is often an indication for wrongly unannotated or wrongly additional annotated nucleotides. Then the sequence file cannot be used for a reliable TIDE analysis. Try if possible to set the right boundary of the decomposition window lower. In case this warning stays, carefully investigate your chromatogram.
wrongly unannotated nucleotide(s)

Low R2 value

A low R2 can be caused when the settings are not optimal or when the sequence quality is not good.

Settings
By default, the decomposition window is set to its maximum size and the Indel size range is set to 10. The settings can be adjusted in advanced settings.

  • Large indels are present in the sample. By default the decomposition is calculated with a maximal indel size of 10. When larger indels are present, they will cannot be modeled, which will result in a low R2. Try to increase the Indel size range to test if this improves the fit (see figure Indel size range)
  • Poor local quality of the sequence trace. Often the end of the sequence is of low quality. This can be observed in the quality plot that shows a high aberrant sequence signal at the end of the sequence trace (see figure Poor quality sequence end). This compromises the decomposition of the sequence trace. Adjust the boundaries of the decomposition window in such a way that it will not overlap with the region that is of low quality.
  • Repetitive regions in the sequence trace. These regions can be observed in the quality plot as a sudden stretch without aberrant nucleotides (see figure Repetitive region). This region might interfere with the decomposition of the sequence trace. Adjust the boundaries of the decomposition window in such a way that it will not overlap with the repetitive sequence part.
Poor sequence quality
Poor sequence quality can be observed in chromatogram (see figure Poor sequence quality). There is more noise present in the data with results in a lower R2.

Figure: Indel size range
maxshift
Figure: Poor quality sequence end
poor local quality
Figure Repetitive region
repetitive region
Figure Poor sequence quality poor sequence quality

Forward and reverse indel spectra are not identical

If the forward and reverse indel spectra don't give the same indel spectra, one of the results is not reliable. Often there is a misannotation in one of the sequence files (see explanation Wrongly annotated nucleotides).

No sgRNA match because there is a mismatch in the control sequence

Sometimes a mismatch occurs in the control sequence at the location of the sgRNA. This will stop the TIDE analysis. In this case, change the guide into identical IUPAC nucleotides as the control sequence.
For example

no sgRNA

Error boundaries of decomposed region are not acceptable

With the used settings the decomposition cannot be performed. This can be caused when the settings are not optimal or when the breaksite is too close to the sequence start or end. Try if possible to set decomposition boundaries further apart or use smaller indel size or use lower the alignment window. If that doesn't help you might have to resequence to perform the TIDE analysis. It can also help to sequence the opposite strand. We advise to sequence a stretch of DNA ~700bp enclosing the designed editing site. The projected break site should be located preferably ~200bp downstream from the sequencing start site.

Poor alignment

When the beginning of the sequence is of poor quality, the alignment function can make a mistake. This can be observed in the quality plot that has high aberrant sequence signal over the whole length of the sequence trace (see figure). The aberrant sequence signal should only increase around the expected cut site (blue dotted line). In case of poor alignment, try to shift the start of sequence read (alignment window) higher or lower. The alignment window can be changed in advance settings. poor alignment