phoenix-cluster / clusters-comparator

Introduction

The Cluster Comparer GUI is a software focused on proteomics mass spectrometry data analysis. The main functions are:

1. Similarity score for spectra(currently supported file formats include MGF file format and *.clustering file format).

2. Basical analusis and visualization of spectra clustering results(currently supported file format include clustering format).

The similarity score algorithm used by the software can be found in the paper(A Pipeline for Differential Proteomics in Unsequenced Species).

*.clustering file is a result file format for MS/MS based spectrum clustering. A detailed description of the *.clustering file format can be found at https://github.com/spectra-cluster/clustering-file-reader.

Analysis and Visualization of *.clustering Files

1. Data Import

Select File > Import Data in the menu bar, select Get Data From Clustering File in the popup window, and the select and submit the *.clustering file to be analysis. Because comparison need two object, you need to enter two files.

2. Result

cluster table and spectrum table

cluster table and spectrum table shows the content of *.clustering file in tables. Each row in the cluster table shows a cluster. spectrum table shows the information of spectra contained in the currently selected cluster.

Supported Operation:

• Click the button switch to display the content of another *.clustering file.
• When clicking different clusters in cluster table, spectrum table, peak map, pie chart and network graph will show the result of the corresponding cluster.
• Change the number of result displayed on each page by page controller, and you can use the controller to jump the specified page number.

peak map

peak map shows all the peaks of specified cluster in the cluster table.

Supported Operation

• Adjust color with color selector
• Zoom in to display the peaks in the rectangular selection area(left mouse button drag selection)

pie chart

pie chart shows the distribution of the spectra contained in the specified cluster in another *.clustering file.

Give an example to illustrate the meaning of the pie chart:

• File I: sample1.clustering

=Cluster=
id=cluster1_1
av_precursor_mz=348.128
av_precursor_intens=1.0
consensus_mz=227.292,288.314,...
consensus_intens=29.8,693.84,191.04,...
SPEC	spec1			348.12817	3.0			0.0
SPEC	spec2			348.12817	3.0			0.0
SPEC	spec3			348.12817	3.0			0.0
SPEC	spec4			348.12817	3.0			0.0

=Cluster=
id=cluster1_2
av_precursor_mz=348.128
av_precursor_intens=1.0
consensus_mz=227.292,288.314,...
consensus_intens=29.8,693.84,...
SPEC	spec5			348.12817	3.0			0.0
SPEC	spec7			348.12817	3.0			0.0
SPEC	spec8			348.12817	3.0			0.0
SPEC	spec9			348.12817	3.0			0.0
SPEC	spec10			348.12817	3.0			0.0

• File II: sample2.Clustering

=Cluster=
id=cluster2_1
av_precursor_mz=348.128
av_precursor_intens=1.0
consensus_mz=227.292,288.314,...
consensus_intens=29.8,693.84,...
SPEC	spec4			348.12817	3.0			0.0
SPEC	spec5			348.12817	3.0			0.0
SPEC	spec6			348.12817	3.0			0.0
SPEC	spec7			348.12817	3.0			0.0
SPEC	spec8			348.12817	3.0			0.0
SPEC	spec9			348.12817	3.0			0.0

=Cluster=
id=cluster2_2
av_precursor_mz=348.128
av_precursor_intens=1.0
consensus_mz=227.292,288.314,...
consensus_intens=29.8,693.84,...
SPEC	spec1			348.12817	3.0			0.0
SPEC	spec2			348.12817	3.0			0.0
SPEC	spec3			348.12817	3.0			0.0

=Cluster=
id=cluster2_3
av_precursor_mz=348.128
av_precursor_intens=1.0
consensus_mz=227.292,288.314,...
consensus_intens=29.8,693.84,191.04,...
SPEC	spec10			348.12817	3.0			0.0
SPEC	spec11			348.12817	3.0			0.0

The cluster1_1 in sample1.clustering contains spectra {spec1, spec2, spec3, spec4}, which are scattered in cluster2_1{spec4} and cluster2_2{spec1, spec2, spec3} in sample2.clustering. This is the meaning of the above pie chart.

Supported Operation:

• Click the block of pie chart to show the comparison of two cluster. The first cluster is the currently specified cluster and the second cluster is from another *.clustering file.

  • Overlapped spectra in two cluster are highlighted.
  • The peak map shows the comparison of peaks of the two cluster and you can use the previously described operation to adjust its display.

• Do Similarity Score

  Click the button Do Similarity Score will perform similarity score between the specified cluster and all clusters displayed in the pie chart(the above example will do similarity score for cluster1_1 vs cluster2_1 and cluster2_2).

  • ID: same as ID in cluster table.

  • Match Rate: the rate of the peaks matched by two spectra under consideration of the fragment tolerance(the number of matched peaks / maxium of tho spectra peaks after processed )

  • Similarity Score: similarity score result Higher similarity score are more similar.

  • Detail: click the buttom Show to display the report of each iteration result.

    • The filtered data has a certain degree of transparency, and the data used for calculation is opaque.

  • Click button View Visualization Result to show visual comparison result, which displays the report of the highest similarity score during two cluster alignment.

network graph

The network graph shows the cluster specified in cluster table is used as the starting point to search the distribution of its related spectra in two *.clustering files. This is different from the search way in pie chart. Specially, we can give an example to explain the meaning of network graph(we still use sample1.clustering and sample2.clustering given above).

The spectra in cluster1_1 are scattered in cluster2_1 and clustering_2_2, but the cluster2_1 also contains spectra {spec5, spec7, spec8, spec9} not in cluster1_1. These spectra are clustered into cluster1_2. And spec10 in cluster1_2 is in cluster2_3. In order to express this relationship in the network graph, we use the node storage cluster and use edges to connect nodes with overlapping spectra. The vertical width of the nodes represents the number of spectra contained in the cluster, and the vertical width of the edges represents the number of spectra shared by two clusters.

Supported Operation:

• Click on the node to display the cluster information
• Click on the edge to show the alignment of the two clusters
• Use the controller to adjust the display of network graph

Analysis and Visualization of MGF Files

1. Data Import

Select File > Import Data in the menu bar, select Get Data From MGF File inthe popup window, and then select and submit the MGF files to be analysis. Because comparison need two object, you need to enter two files.

2. Result

MGF table

MGF table shows the content of MGF file in tables. Each row in MGF table shows a spectrun.

Supported Operation

• Click on the different spectra in table, Similarity Score Table and Peak Map will change accordingly

Similarity Score table

Similarity Score table shows the similarity score between the current spectrum and all the spectra in another file.

• ID: same as Title of MGF table.

Other properties are the same as description in Do Similarity Score

Next Version

• Modify the organizational form of the relevant code to decouple
• Modify GUI view.