BRCA Heatmap

import CanDI.candi as can
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid1 import make_axes_locatable

Cancer Object Instantiation

We’re interested in cross referencing some data in breast and ovarian cancer so instantiate cancer objects as follows. To double check the object instantiation I check the length of the depmap_id vectors. This lets me know we’re able to index other datasets correctly

ov = can.Cancer("Ovarian Cancer")
br = can.Cancer("Breast Cancer")

#Number of Ovarian Cell lines
print(len(ov.depmap_ids))
#Number of Breast Cell Lines
print(len(br.depmap_ids))

74
83

Subsetting by mutation status

Explicitly load mutations into memory.This only needs to be done once You will be done prompted to load a given dataset if using operations that act on that dataset and it is not in memory.

can.data.load("mutations")

	gene	Entrez_Gene_Id	NCBI_Build	Chromosome	Start_position	End_position	Strand	Variant_Classification	Variant_Type	Reference_Allele	...	isCOSMIChotspot	COSMIChsCnt	ExAC_AF	Variant_annotation	CGA_WES_AC	HC_AC	RD_AC	RNAseq_AC	SangerWES_AC	WGS_AC
0	VPS13D	55187	37	1	12359347	12359347	+	Nonsense_Mutation	SNP	C	...	False	0.0	NaN	damaging	34:213	NaN	NaN	NaN	34:221	NaN
1	AADACL4	343066	37	1	12726308	12726322	+	In_Frame_Del	DEL	CTGGCGTGACGCCAT	...	False	3.0	NaN	other non-conserving	57:141	NaN	NaN	NaN	9:0	28:32
2	IFNLR1	163702	37	1	24484172	24484172	+	Silent	SNP	G	...	False	0.0	NaN	silent	118:0	NaN	NaN	10:0	118:0	18:0
3	TMEM57	55219	37	1	25785018	25785019	+	Frame_Shift_Ins	INS	-	...	False	0.0	NaN	damaging	NaN	NaN	NaN	6:28	NaN	NaN
4	ZSCAN20	7579	37	1	33954141	33954141	+	Missense_Mutation	SNP	T	...	False	0.0	NaN	other non-conserving	28:62	NaN	NaN	NaN	27:61	NaN
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
1230240	SLC39A4	55630	37	8	145641991	145641991	+	Silent	SNP	C	...	False	0.0	0.000017	silent	52:23	NaN	NaN	NaN	NaN	NaN
1230241	TAL2	6887	37	9	108424778	108424778	+	Start_Codon_SNP	SNP	A	...	False	0.0	NaN	damaging	27:0	NaN	NaN	NaN	NaN	NaN
1230242	TRO	7216	37	X	54955098	54955098	+	Silent	SNP	C	...	False	0.0	NaN	silent	5:16	NaN	NaN	NaN	NaN	NaN
1230243	USP51	158880	37	X	55514703	55514703	+	Missense_Mutation	SNP	G	...	False	0.0	NaN	other non-conserving	23:0	NaN	NaN	NaN	NaN	NaN
1230244	C1GALT1C1	29071	37	X	119760406	119760406	+	Missense_Mutation	SNP	T	...	False	0.0	NaN	other non-conserving	28:0	NaN	NaN	NaN	NaN	NaN

1230245 rows × 32 columns

I want to look at BRCA1 mutations in these types of cancers. I start by using the mutated function to identify ovarian and breast cancer cell lines with BRCA1 mutations. A cancer object’s mutated method’s default behavior is to output a list of depmap ids corresponding to celllines containing any mutation within the given genes. I then instantiate CellLineCluster objects of exclusively mutated or wild type cell lines for both breast and ovarian cancer. This makes comparing these cell lines easier.

ov_mt_list = ov.mutated(["BRCA1"]) #List of depmap_ids
br_mt_list = br.mutated(["BRCA1"]) #list of depmap_ids

ov_mt = can.CellLineCluster(ov_mt_list) #CellLineCluster obj
br_mt = can.CellLineCluster(br_mt_list)


print("Depmap_ids attribute should be the same as the list used to instantiate the CellLineCluster object\n")
print(ov_mt.depmap_ids == ov_mt_list)

#CellLineCluster ojbect must be instantiated with a mutable sequence
#I use set operations to get wild type cell line ids and convert to a list
ov_wt_list = list(set(ov.depmap_ids) - set(ov_mt_list))
br_wt_list = list(set(br.depmap_ids) - set(br_mt_list))

ov_wt = can.CellLineCluster(ov_wt_list)
br_wt = can.CellLineCluster(br_wt_list)
print(ov_wt.depmap_ids == ov_wt_list)

Depmap_ids attribute should be the same as the list used to instantiate the CellLineCluster object

True
True

Cross Referencing Mutation and Gene Knockout Data

I’m interested in how the mutation status of BRCA1 effects a cancer’s dependency on the fanconi anemia genes. To visualize this relationship I am going to make a heatmap of fanconi anemia genes sorting the cell lines by their BRCA1 mutation status. The following cell defines a function that plots a heatmap of the gene effect of the fanconi anemia genes separating them by the BRCA1 mutation status of a given cell line.

def gene_effect_heatmap(obj1, obj2, genes, name = None):
    
    #Make Figure appropriate size, dpi, and font
    plt.rcParams.update({"figure.figsize": (16, 6),
                        "savefig.dpi": 300,
                        "font.size": 12
                        })
    
    #One figure with one subplot
    fig, ax = plt.subplots(1,1)
    
    #Construcat matrix to make heatmap and cell line labels
    data = pd.concat([obj1.effect_of(genes), obj2.effect_of(genes)], axis=1)
    names = can.data.cell_lines.loc[data.columns, "cell_line_name"]
        
    # We want to show all ticks...
    ax.set_xticks(np.arange(len(names)))
    ax.set_yticks(np.arange(len(genes)))
    # ... and label them with the respective list entries
    ax.set_xticklabels(names)
    ax.set_yticklabels(genes)
    
    #make heatmap
    im = ax.imshow(data, cmap="RdBu")
    
    #Make colorbar scale to axis
    divider = make_axes_locatable(ax)
    cax = divider.append_axes("right", size="5%", pad=0.1)
    cbar = ax.figure.colorbar(im, ax = ax, cax = cax)
    cbar.ax.set_ylabel("Gene Effect", rotation=-90, va="bottom")
    
    #Draw Dividing line btween mutant and
    ax.axvline(x=obj1.gene_effect.shape[1] - 0.5, c = "black", linewidth = 3)
    plt.setp(ax.get_xticklabels(), rotation=-90, ha="left", va="center",
         rotation_mode="anchor")
    plt.tight_layout()
    plt.show()
    
    if name:
        fig.savefig(name, dpi=300)

Fanconi Anemia Genes Knockout Effect in Ovarian Cancer

BRCA1 Mutant Left of Vertical Line

genes = ["FANC" + i for i in ["A", "C", "I", "M", "D2", "F", "E"]]
gene_effect_heatmap(ov_mt, ov_wt, genes, name = None)

Load Complete

Fanconi Anemia Genes Knockout Effect in Breast Cancer

BRCA1 Mutant Left of Vertical Line

gene_effect_heatmap(br_mt, br_wt, genes, name = None)