The Peak calling task is used to detect enriched genomic regions on reads generated from nucleic acid enrichment experiments such as ChIP-seq, DNase-seq, and MeDIP-seq. experiments. Partek^® Flow^® provides the widely used method of MACS2-model-based analysis¹ (http://liulab.dfci.harvard.edu/MACS/) to find peaks. It can be performed with or without control sample.

MACS2 dialog

Selecting MACS2 from the context sensitive menu will bring up the MACS2 task dialog. The interface will appear differently depending on the input aligned data node and whether there are sample attributes available in the Data tab.

If the selected aligned data node was imported, the reference assembly used during data alignment needs to be specified. Choose the Assembly from the drop-down list within the MACS2 dialogue (Figure 1). If the selected aligned data node was generated by Partek Flow, this option will not appear.
 

The Effective genome size must be configured prior to running the peak caller. It refers to the size of the genomic regions that are actually mappable. This size is smaller than the actual size of an organism's whole genome because of the presence of repetitive features. They are typically about 70%-90% of the whole genome. The MACS2 authors¹  have recommended presets available for four different species. Select from the drop-down menu the preset that best describes the genome you are working with. They are as follows:

• Human (Homo sapiens) – 2.7 x 10⁹
• Mouse (Mus musculus) – 1.87 x 10⁹
• C. elegans – 9 x10⁷
• Fruitfly (Drosophila melongaster) – 1.2 x 10⁸

If none of these presets match your genome of interest, select Other... Then enter the effective genome size (Figure 2). The values are in base pairs (bps). Consult the MACS documentation for guidance on selecting the best effective genome size for your experimentWhile processing 10x Chromium Single Cell Multiome ATAC + Gene Expression sequencing data via ‘cellranger-arc count’ pipeline, Feature linkages analysis is performed as pairs of genomic features, such as peaks and genes, that have significant correlation in signals across cells. Because it provides a basis for inferring enhancer-gene targeting relationships and constructing transcriptional networks. The features with strong linkages are considered to be “co-expressed” and enriched for a shared regulatory mechanism.

Partek Flow provides the opportunity to our users to explore the linkage relationships among different features including peaks and genes, peaks and peaks, and genes and genes. A tab-delimited file containing information of feature linkages inferred from Flow Cell Ranger - ATAC task will be loaded into Integrative Genome Viewer (IGV)[1] for exploration if Feature linkage analysis task has been completed successfully.

Running Feature linkage analysis

To run Feature linkage analysis task (Figure 1),

• Click one datanode that has both features of ATAC and gene expression;
• Click the Feature linkage analysis task under Peak analysis section in the toolbox;
• Click the Finish button to complete the submission.

For data where no sample attributes are specified, the peak detection pairs need to be manually defined. In the example in Figure 1, there only two samples. Under the Define pairs section, the left panel lists all the sample names uploaded to the project (H3K27 and Mock).  Add one pair at a time by dragging the corresponding samples to either the IP panel on the top-right or the Control panel on the bottom-right. If no control samples are present in the experiment, leave the Control panel blank. If more than one ChIP or Control samples are added, the samples will be combined (or pooled) during the analysis. After defining a pair, click the Add pair button.

...

There will be no inputs needed if the FASTQ is converted to counts matrix within Flow. However, if users processed the FASTQ files outside of Partek, and imported the counts matrix into Flow later. The feature_linkage.bedpe file in outs/analysis/feature_linkage from Cell Ranger output will be needed for each sample (Figure 2) to complete the analysis.

When running the MACS2 task, sample attributes will be used to define the multiple pairs (Figure 4). There is an IP-Input pair for each time point, so the Pair attribute is the Time attribute. The Control attribute is the attribute that differentiates between the Input and IP groups, and in this example, it is the ChIP attribute.  Finally, the Control term is labeled as Input in the example. 

...

Task report

A new datanode will be displayed as the task is finished. Double click on the output datanode, Flow will bring you to the IGV browser where you could explore the “co-expressed”feature pairs (Figure 3).

If multiple pairs are added in the Pairs table, the peak detection is performed on each pair independently.

Peaks report

In the task report, each pair will generate a list of peaks displayed in a table (Figure 6). Use the drop down menu next to Peaks detected for... to select the pair.

The feature_linkage.bedpe file[2] outputted from Cell Ranger pipeline is available in task report as a table. In the report table, each row is a region of a peak and includes the following information:

• Absolute summit: base pair location of peak summit
• Pileup: pileup height at peak summit
• -log10(pvalue): negative log10 pvalue for the peak summit
• Fold enrichment: fold enrichment for the peak summit against random Poisson distribution with local lambda
• -log10(qvalue): negative log10 qvalue at peak summit
• a peak name generated by the MACS2 algorithm

Click the browse to peak button (Image Removed) to invoke chromosome view and zoom into that location.

Click the Download button at the lower-right corner to download the peaks in a text file.

References

...

• Sample name: name for each sample
• Chromosome 1: the name of the chromosome on which the first feature exists.
• Start 1: the starting position of the first feature on that chromosome.
• End 1: the ending position of the first feature on that chromosome.
• Chromosome 2: the name of the chromosome on which the second feature exists.
• Start 2: the starting position of the second feature on that chromosome.
• End 2: the ending position of the second feature on that chromosome.
• Name: the name of the linkage features with the format of <name1><name2>, in which name1 and name2 are based on gene symbol or peak annotation. 
• Score: linkage correlation, ranging from -1 to 1.
• Strand 1: all set to ".".
• Strand 2: all set to ".".
• Significance: linkage significance: -log10 (p-value) after multiple testing correction (FDR, false discovery rate). Capped at 299.
• Distance: distance in base pairs from feature 2 to feature 1.
• Linkage type: can be "peak-peak", "gene-peak" or "peak-gene" depending on the type of gene or peak for feature 1 and feature 2.

Filter feature linkage task

To filter out and visualize only the linkages that users are interested in is also made possible through the Filter feature linkages task in Flow (Figure 4). Users are able to download the .bedpe file from Flow and explore them via their stand-alone IGV.

References

1. https://software.broadinstitute.org/software/igv/
2. https://support.10xgenomics.com/single-cell-multiome-atac-gex/software/pipelines/latest/output/analysis

...