#!/usr/bin/env python
# coding: utf-8

# # Merged PTM and Expression SCLC Cluster
# This notebook will investigate the cluster of up-regulated PTMs and genes in SCLC lung cancer cell lines. This cluster was isolated and saved in the notebook: [CST_Data_Viz.ipynb](http://nbviewer.jupyter.org/github/MaayanLab/CST_Lung_Cancer_Viz/blob/master/notebooks/CST_Data_Viz.ipynb#SCLC_Cluster).

# In[1]:


from clustergrammer_widget import *
net = Network(clustergrammer_widget)


# In[2]:


# load data
net.load_file('histology_clusters/merge_sclc.txt')
merge_sclc = net.export_df()

# manually set category colors for rows and columns
net.set_cat_color('row', 1, 'Data-Type: phospho', 'red')
net.set_cat_color('row', 1, 'Data-Type: Rme1', 'purple')
net.set_cat_color('row', 1, 'Data-Type: AcK', 'blue')
net.set_cat_color('row', 1, 'Data-Type: Kme1', 'grey')
net.set_cat_color('row', 1, 'Data-Type: Exp', 'yellow')
net.set_cat_color('col', 1, 'Histology: SCLC', 'red')
net.set_cat_color('col', 1, 'Histology: NSCLC', 'blue')
net.set_cat_color('col', 2, 'Sub-Histology: SCLC', 'red')
net.set_cat_color('col', 2, 'Sub-Histology: NSCLC', 'blue')
net.set_cat_color('col', 2, 'Sub-Histology: squamous_cell_carcinoma', 'yellow')
net.set_cat_color('col', 2, 'Sub-Histology: bronchioloalveolar_adenocarcinoma', 'orange')
net.set_cat_color('col', 2, 'Sub-Histology: adenocarcinoma', 'grey')


# # Cluster of Up-regulated PTMs and Genes in SCLC Cell Lines
# Below we will visualize the SCLC cluster. We will highlight a few of the interesting genes/proteins and sub-clusters found in this cluster of up-regulated PTMs and gene expression levels data in SCLC cell lines. 

# In[3]:


net.cluster(views=[])
net.widget()


# We can see that the cluster is composed of roughly for smaller clusters that are generally up-reglated in SCLC cell lines and down-regulated in NSCLC cell lines. We can also see that the cluster is composed of a mixture of phosphorylation, expression, methylation and acetylation. 
# 
# ### NKX2-1 and SOX2 Cluster
# NKX2-1 is a transcription factor that is known to have a role in lung development and has been used as a biomarker in lung cancer  ([Yang et al. 2012](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3494024/)). In this cluster we find NKX2-1 expression and NKX2-1 methylation (R121) are both present and cluster closely together. The expression and methylation data are from independent sources and this NKX2-1 co-clustering demonstrate broad agreement between the two datasets. The data imply that variability in NKX2-1 methylation may be the result of variability in its expression level. Below is a screenshot of the immediate cluster of phophorylation, expression, and methylation including NKX2-1:
# 
# ![NKX2-1_cluster.png](img/NKX2-1_cluster_new.png)  
# 
# NKX2-1 levels are known to be generally high in SCLC cell lines which broadly agrees with our results. NKX2-1 also clusters with several other lung associated genes including: 
# 
# * STFA3 is a lung associated protein with possible immune functions ([Schict et al. 2014](https://www.ncbi.nlm.nih.gov/pubmed/24743970)) 
# * SOX2 is known to be involved in lung cancer ([Karachaliou et al. 2013](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4367598/)) and is also thouhgt to interact with NKX2-1 ([Ferri et al. 2013](https://www.ncbi.nlm.nih.gov/pubmed/23444355)) 
# * GPNMB is thought to be involved in lung cancer ([Oyewumi et al. 2016](https://www.ncbi.nlm.nih.gov/pubmed/26883195)) 
# * ID4 has been found to be associated with lung cancer metastisis ([Pan et al. 2015](http://cancerres.aacrjournals.org/content/75/15_Supplement/1431)) 
# * PEG10 is thought to play a role in lung cancer proliferation ([Deng et al 2014](https://www.ncbi.nlm.nih.gov/pubmed/25199998))
# * CTHRC1 overexpression is though to be associated with NSCLC tumor aggressiveness ([Ke et at. 2014](https://www.ncbi.nlm.nih.gov/pubmed/25238260)).
# 
# This cluster also includes MARCKSL1 (MARCKS-like 1) phosphorylations: S71, T148, and S151. MARCKSL1 is thought to play a role in cytoskeletal reglation [see Gene Cards](http://www.genecards.org/cgi-bin/carddisp.pl?gene=MARCKSL1). This implies that phosphorylation of MARCKSL1 phosphorylation may play a role in lung development based on its co-clustering with NKX2-1 expression/methylation.
# 
# ### Cytoskeletal and Extracellular Matrix Proteins
# Several cytoskeletal and extracellular matrix (ECM) associated genes/proteins appear in this cluster. This highlights the importance cytoskeletal/ECM function in SCLC lung cancer.
# * Actin: ACTB is acetylated at K328, K61. 
# * Tubulin: TUBA1A exp, TUBA1A Ack K394, TUBB phos S115, 
# * Collagens: COL11A1, COL1A2, COL4A1, COL6A3 (expression and correlation)
# * Rac GTPase: RACGAP1 (Rac GTPase activiting protein 1) is phosphorylated at T606 (regulates actin)
# * Myosin: MYH9 acetylated, MYH10 acetylated, MYH11 acetylated, MYLK phosphorylated, MYO9A phos
# * Stathmin a microtubule depolymerizer: STMN1 Ack K53, STMN2-exp (stathmin). Stathmin expression is thought to be a poor prognostic biomarker in NSCLC ([Nie et al. 2015](https://www.ncbi.nlm.nih.gov/pubmed/25384122)). Stathmin 2 expression and Stathmin 1 acetylation are similar. 
# * Microtubule associated proteins: MAP2 exp: thought to be involved in MT assembly and neurogenesis. MAP7, primarily expressed in epithelial. DCX, doublecortin, microtubule binding and involved in the development of the cortex.

# # Gene Ontology Biological Process 2015
# Here we are pre-calculating enrichment for biological processes from the Gene Ontology resource using the ``enrichrgram`` method (see [Clustergrammer-PY's API](http://clustergrammer.readthedocs.io/clustergrammer_py.html#clustergrammer-py-api) for more information). This will help us understand the broad biological processes occurring in this cluster of up-regulated PTMs and Genes.

# In[4]:


net.enrichrgram('GO_Biological_Process_2015')
net.cluster(views=[])
net.widget()


# ### mRNA Processing and Gene Expression 
# We see enrichment for mRNA processing, splicing, and gene expression. We also see similar results with KEGG 2016 enrichment, which can be run in this notebook with the [Enrichrgram button](http://clustergrammer.readthedocs.io/biology_specific_features.html#enrichment-analysis). 
# 
# ### Neuronal Sub-Cluster
# Genes with the above ontologies are mainly distributed at the top and bottom sub-clusters - note that the largest middle sub-cluster with highly up-regulated values has relatively few genes with this association. We can investigate the biological processes in the middle sub-cluster by clicking the [dedrogram crop button](http://clustergrammer.readthedocs.io/interacting_with_viz.html#interactive-dendrogram) and re-running the enrihment analysis using the front-end [Enrichrgram button](http://clustergrammer.readthedocs.io/biology_specific_features.html#enrichment-analysis). 
# 
# Doing so reveals that this cluster is also enriched for gene expression and mRNA processing but also for several neuronal processes including: neuron projection, axon guidance and neuron projection morphology. This agrees with the neuronal characteristics of SCLC cell lines (see [Onganar et al 2005](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2361510/)). 

# # Disease Perturbations from GEO Up
# Using the 'Disease Perturbations from GEO Up' [Enrichr](http://amp.pharm.mssm.edu/Enrichr/) library we can find diseases that have up-regulated genes that are similar to our up-regulated genes/proteins. This can help us understand the disease-associations of the differntially expressed genes and proteins with differentially reglated PTMs. 

# In[5]:


net.enrichrgram('Disease_Perturbations_from_GEO_up')
net.cluster(views=[])
net.widget()


# Above, we see that up-regulated genes and PTMs in SCLC are similar to up-regulated genes in the following diseases: oligodendroglioma, multiple sclerosis, and other cancers. 

# # MGI Mammalian Phenotype
# To find phenotypes associated with our genes/proteins, we can use the 'MGI Phenotype Level 4' Enrichr library. This library will return phenotypes that are associated with gene knockout in mice. This can give us a less biased overview of gene/phenotype relationships. 

# In[6]:


net.enrichrgram('MGI_Mammalian_Phenotype_Level_4')
net.cluster(views=[])
net.widget()


# ### Perinatal Lethality and Neuronal Abnormalities
# Above we see that our genes/proteins are enriched for the following broad phenotypes: 
# * perinatal lethality (prenatal, preweaning, and postnatal lethality) 
# * neuronal abnormalities (neuron morphology, brain morphology, spinal cord morphology, and nervous system abnormalities). 
# The enrichment for neuronal abnormality phenotypes agrees with the neuronal enrichment we obtained above with Gene Ontology Biological Process and with our prior knowledge of SCLC characteristics. 

# In[ ]: