BinderSpace Tutorial

Data information: The input data is peptide sequence data that bind to Bcl-2 family of proteins, the dataset id is: GSE118147. For more details, refer to this link GSE118147.

• positive sequence: GSM3319530_m1r.csv' (peptide sequences that bind to Mcl-1)
• negative sequence: GSM3319530_x100r.csv and GSM3319530_x1.csv (peptide sequences that bind to Bcl-xL)

1. Installing

Downloading the motif_search.py, VisualizeMotif.ipynb and put it in the same folder with your data.

2. Run motif_search.py

1. Prepare input file: #####The input file is a CSV file contains DNA or peptide sequences. The sequences column header should be “Sequence”. Here we use 22_nt peptide sequences (3480 rows) from Z-keating dataset, and we also get 1470 negative sequences from peptide that bind to other proteins (pos_seqs.csv, neg_seqs.csv). The sequences should have same length and the sequences should not contain letters beyond 22 amino acids or 4 DNA bases.

2. Run motif search: #####Parameter setting: no gap, motif minimal length is 8, minmal frequency is 0.5% * total number of positive sequences, use 20 processors #####command line: python motif_search.py -i pos_seqs.csv -n neg_seqs.csv -o keating -m 8 -f 0.005 -g 2 -a 2

3. Output: #####The output keating_length*.csv files contains 8 columns: motifs, frequency of this motif occurs in positive sequences (positive_occurrence %), frequency of this motif occurs in negative sequences, percentage of the positive occurrence, percentage of the negative occurrence, index of positive sequences, index of negative sequences

4. Other command options: run “python MotifSearch.py“ to see the parameter options

• -i posfile -- the file with positive sequences (REQUIRED)

• -n negfile -- the file with negative sequences
• -o motif_file --the name of the output file (default: 'motifs')
• -f minfreq -- the fraction of the minimal motif occurence in positive sequences (default: f = 0.001, minfreq = f * (the total number of positive sequences))
• -m minlength --the minimal length of the motifs (default: 3)
• -l maxlength --the maximal length of the motifs (default: the length of the positive sequences)
• -c category --choose from 'dna' or 'amino_acids',(default: 'amino_acids')
• -g maxgaps --maximal number of gaps (default: 0)
• -a maxgaplength --maximal gap length (default: 1)
• -p processor --number of processors used for running the program (default: 20) italicized text

##3. PCA/tSNE visualization for one of the motifs (this part only works for dataset that has negative control)

Open VisualizeMotif.ipynb by jupyter notenook. Choose one motif file (‘keating_length 11.csv’) and from that motif file choose one motif ('LRIGDEDAY'), and we will check the locations of those sequences that contain that motif in PCA and t-SNE space.
Input:

pos_seq = ‘pos_seqs.csv’

Output: