app.py

import os
import numpy as np
import pandas as pd
import streamlit as st

import scanpy as sc

#import mpld3
import matplotlib.pyplot as plt
#import seaborn as sns


#import streamlit.components.v1 as components

#from IPython.display import Markdown as md

from functions import pathway_analyses

# SMALL_SIZE = 2
# MEDIUM_SIZE = 2
# BIGGER_SIZE = 2

# plt.rc('font', size=SMALL_SIZE)          # controls default text sizes
# plt.rc('axes', titlesize=SMALL_SIZE)     # fontsize of the axes title
# plt.rc('axes', labelsize=MEDIUM_SIZE)    # fontsize of the x and y labels
# plt.rc('xtick', labelsize=SMALL_SIZE)    # fontsize of the tick labels
# plt.rc('ytick', labelsize=SMALL_SIZE)    # fontsize of the tick labels
# plt.rc('legend', fontsize=SMALL_SIZE)    # legend fontsize
# plt.rc('figure', titlesize=BIGGER_SIZE)  # fontsize of the figure title

sc.settings.set_figure_params(dpi=80, facecolor='white')

#disable st.pyplot warning
st.set_page_config(layout="wide") 
st.markdown(
    """
<style>
.streamlit-expanderHeader {
    font-size: x-large;
}
</style>
""",
    unsafe_allow_html=True,
)
m=st.markdown("""

<style>

div.stTitle {

font-size:40px;

}

</style>"""
,unsafe_allow_html=True)

st.set_option('deprecation.showPyplotGlobalUse', False)

#load Data
cwd=os.getcwd()+'/'#+'data/'
@st.cache_data
def get_data():
    if 'adata_annot' not in st.session_state:
        adata_annot = sc.read_h5ad(cwd+'multiregion_brainaging_annotated.h5ad')
        st.session_state['adata_annot'] = adata_annot
        if 'genes_list' not in st.session_state:        
            genes=adata_annot.var.index
            #genes_list=sorted(genes.unique())
            st.session_state['genes_list'] = sorted(genes.unique())
        if 'cell_type' not in st.session_state:
            #cell_type=diff_fdr[diff_fdr.type=='cell_type']['tissue']
            #cell_type=sorted(cell_type.unique())
            anno=adata_annot.obs.new_anno
            #cell_type=sorted(anno.unique())
            st.session_state['cell_type'] = sorted(anno.unique())
        if 'broad_type' not in st.session_state:        
            broad_celltype=adata_annot.obs.broad_celltype
            #broad_type=sorted(broad_type.unique())
            st.session_state['broad_type'] = sorted(broad_celltype.unique())
            
    #Also load Go Terms
    if 'go_table' not in st.session_state:        
        bp = pathway_analyses.read_pathways('pathway_databases/GO_Biological_Process_2021.txt')

        # cy = pathway_analyses.read_pathways('pathway_databases/HumanCyc_2016.txt')
        # ke = pathway_analyses.read_pathways('pathway_databases/KEGG_2019_Human.txt')
        # re = pathway_analyses.read_pathways('pathway_databases/Reactome_2016.txt')

        # all_paths = pd.concat([bp, cy, ke, re], join='outer', axis=0, ignore_index=True)
        # all_paths.set_index(0, inplace=True)
        # all_paths.fillna("", inplace=True)
        # all_paths_dict = all_paths.to_dict(orient='index')


        go_bp_paths = bp.set_index(0)
        go_bp_paths.fillna("", inplace=True)
        go_bp_paths_dict = go_bp_paths.to_dict(orient='index')


        gene_set_by_path = {key: [val for val in value.values() if val != ""] for key, value in go_bp_paths_dict.items()}
        gene_set_by_path = pd.DataFrame.from_dict(gene_set_by_path, orient='index').transpose()     
        st.session_state['path_ways']=gene_set_by_path.columns
        st.session_state['go_table']=gene_set_by_path
#done load Data



#st.title('Single nuclei atlas of human aging in brain regions')
st.title('Brain Age Browser')

#genes_list,adata_annot=get_data()

get_data()

tab1, tab2,readme = st.tabs(["Gene Expression by CellType", "Age associations for multiple genes", "README"])
data = np.random.randn(10, 1)
with tab1:
    with st.form(key='columns_in_form'):
        #c1, c2, c3 = st.columns([4,4,2])
        c1, c2 = st.columns(2)
        with c1:
            selected_gene = st.selectbox(
            'Please select a gene',
            st.session_state['genes_list'])
        with c2:
            selected_celltype = st.selectbox(
            'Please select CellType',
            st.session_state['cell_type']
            )
        # with c3:
        #     plot_choice = st.checkbox(
        #     "Which Plots",
        #     ('Gene','Old/Young'))
            
            

        Updated=st.form_submit_button(label = 'Go')
    if not isinstance(selected_gene, type(None)) and not isinstance(selected_celltype, type(None)) and Updated:


        # fig11, axx1 = plt.subplots()
        # sc.pl.umap(st.session_state['adata_annot'], color='new_anno', title='', legend_loc='on data',legend_fontsize='4', frameon=False,show=False, ax=axx1)
        # st.pyplot(plt.gcf().set_size_inches(4, 4))     

        col1,col2= st.columns([1,1])
        with col1:
            fig11, axx11 = plt.subplots(figsize=(5,5))
            sc.pl.umap(st.session_state['adata_annot'], color='new_anno', title='', legend_loc='on data',legend_fontsize='8', frameon=False,show=False, ax=axx11)
            st.pyplot(fig11)   
                 
        with col2:
            fig12, axx12 = plt.subplots(figsize=(5,5))
            #sc.pl.umap(st.session_state['adata_annot'], color='new_anno', title='', legend_loc='on data', frameon=False,show=False, ax=axx2)
            sc.pl.umap(st.session_state['adata_annot'], color=selected_gene, title=selected_gene, legend_loc='best', frameon=False,show=False,legend_fontsize='xx-small', ax=axx12)#,vmax='p99')
            #plt.xticks(rotation = 45)
            st.pyplot(fig12)        

        #Subset Younv and Old
        adata_Young = st.session_state['adata_annot'][st.session_state['adata_annot'].obs['Age_group']=='young']
        adata_Old = st.session_state['adata_annot'][st.session_state['adata_annot'].obs['Age_group']=='old']
        
        #Young/Old  but for cell_type
        adata_YoungAst = adata_Young[adata_Young.obs['new_anno']==selected_celltype]
        adata_OldAst = adata_Old[adata_Old.obs['new_anno']==selected_celltype]
        
        # # #Young/Old  but for cell_type
        # # adata_YoungAst = adata_Young[adata_Young.obs['broad_celltype']==selected_celltype]
        # # adata_OldAst = adata_Old[adata_Old.obs['broad_celltype']==selected_celltype]
        
        #Young
        dot_size=.05
        col1,col2= st.columns([1,1])
        
        with col1:
            
            #st.markdown('<div style="text-align: center;">**Young**</div>', unsafe_allow_html=True)
            str_title='Young: '+selected_gene
            #st.markdown("<h3 style='text-align: center; color: red;'>str_title</h3>", unsafe_allow_html=True)
            st.markdown("# {} ".format(str_title))#,align_text='center')
            #md("# {} ".format(str_title))

            fig21, axx21 = plt.subplots(figsize=(1,1))
            #sc.pl.umap(adata_Young, color=selected_gene, title="Young: "+selected_gene, legend_loc='right margin', color_map='viridis',frameon=False,show=False,size=dot_size, legend_fontsize='4',colorbar_loc=None,ax=axx21)  
            sc.pl.umap(adata_Young, color=selected_gene, title="", legend_loc='right margin', color_map='viridis',frameon=False,show=False,size=dot_size, legend_fontsize='xx-small',colorbar_loc=None,ax=axx21)  
            #st.pyplot(fig21)  
            st.pyplot(plt.gcf())        
        with col2:
            str_title='Young: '+selected_gene+" ("+selected_celltype+")"
            st.markdown("# {} ".format(str_title))#,align_text='center')
            fig22, axx22 = plt.subplots(figsize=(1,1))
            #sc.pl.umap(st.session_state['adata_annot'], color='new_anno', title='', legend_loc='on data', frameon=False,show=False, ax=axx2)
            #sc.pl.umap(st.session_state['adata_annot'], color=selected_gene, title=selected_gene, legend_loc='best', frameon=False,show=False, ax=axx2)#,vmax='p99')
            sc.pl.umap(adata_YoungAst, color=selected_gene, title="", legend_loc='right margin', color_map='viridis', frameon=False,show=False,size=dot_size,legend_fontsize='xx-small',colorbar_loc=None, ax=axx22)  
            #sc.pl.umap(adata_Old, color=selected_gene, title="Old: "+selected_gene, legend_loc='right margin', color_map='viridis', frameon=False,show=False, ax=axx22)  
            #plt.xticks(rotation = 45)
            #st.pyplot(fig22)        
            st.pyplot(plt.gcf())  
        #Old
        col1,col2= st.columns([1,1])
        with col1:
            str_title='Old: '+selected_gene+" ("+selected_celltype+")"
            st.markdown("# {} ".format(str_title))#,align_text='center')
            
            fig31, axx31 = plt.subplots(figsize=(1,1))
            #sc.pl.umap(st.session_state['adata_annot'], color='new_anno', title='', legend_loc='on data',legend_fontsize='8', frameon=False,show=False, ax=axx1)
            sc.pl.umap(adata_Old, color=selected_gene, title="", legend_loc='right margin', color_map='viridis', frameon=False,show=False,size=dot_size,legend_fontsize='xx-small', colorbar_loc="bottom",ax=axx31)  
            st.pyplot(fig31)        
        with col2:
            str_title='Old: '+selected_gene+" ("+selected_celltype+")"
            st.markdown("# {} ".format(str_title))#,align_text='center')
            
            fig32, axx32 = plt.subplots(figsize=(1,1))
            #sc.pl.umap(st.session_state['adata_annot'], color='new_anno', title='', legend_loc='on data', frameon=False,show=False, ax=axx2)
            #sc.pl.umap(st.session_state['adata_annot'], color=selected_gene, title=selected_gene, legend_loc='best', frameon=False,show=False, ax=axx2)#,vmax='p99')
            sc.pl.umap(adata_OldAst, color=selected_gene, title="", legend_loc='right margin', color_map='viridis', frameon=False,show=False,size=dot_size,legend_fontsize='xx-small', colorbar_loc="bottom",ax=axx32)  
            #plt.xticks(rotation = 45)
            st.pyplot(fig32)        
        
        # fig, ax = plt.subplots(3,2)
        # sc.pl.umap(st.session_state['adata_annot'], color='new_anno', title='', legend_loc='on data', frameon=False,show=False, ax=ax[0,0])
        # sc.pl.umap(st.session_state['adata_annot'], color=selected_gene, title=selected_gene, legend_loc='best', frameon=False,show=False, ax=ax[0,1],vmax='p99')
        
        # #Subset Younv and Old
        # adata_Young = st.session_state['adata_annot'][st.session_state['adata_annot'].obs['Age_group']=='young']
        # adata_Old = st.session_state['adata_annot'][st.session_state['adata_annot'].obs['Age_group']=='old']
        # sc.pl.umap(adata_Young, color=selected_gene, title="Young: "+selected_gene, legend_loc='right margin', color_map='viridis',frameon=False,show=False, ax=ax[1,0])  
        # sc.pl.umap(adata_Old, color=selected_gene, title="Old: "+selected_gene, legend_loc='right margin', color_map='viridis', frameon=False,show=False, ax=ax[2,0])  
        
        # # #Young/Old  but for cell_type
        # # adata_YoungAst = adata_Young[adata_Young.obs['broad_celltype']==selected_celltype]
        # # adata_OldAst = adata_Old[adata_Old.obs['broad_celltype']==selected_celltype]

        # #Young/Old  but for cell_type
        # adata_YoungAst = adata_Young[adata_Young.obs['new_anno']==selected_celltype]
        # adata_OldAst = adata_Old[adata_Old.obs['new_anno']==selected_celltype]
        
        
        # sc.pl.umap(adata_YoungAst, color=selected_gene, title=selected_celltype, legend_loc='right margin', color_map='viridis', frameon=False,show=False, ax=ax[1,1])  
        # sc.pl.umap(adata_OldAst, color=selected_gene, title=selected_celltype, legend_loc='right margin', color_map='viridis', frameon=False,show=False, ax=ax[2,1])  
        
        # #sc.pl.umap(st.session_state['adata_annot'], color='Brain_region', title='Brain Region', legend_loc='right margin', frameon=False,show=False, ax=ax[1,1])  
        # #sc.pl.umap(st.session_state['adata_annot'], color='Age_group', title='Age Group', legend_loc='right margin', frameon=False,show=False, ax=ax[2,0])  
        # #sc.pl.umap(st.session_state['adata_annot'], color=selected_celltype, title=selected_celltype, legend_loc='on data', frameon=False,show=False, ax=ax[2,1])  
        
        
        # st.pyplot(plt.gcf().set_size_inches(15, 30))
                


with tab2:
    with st.form(key='multiselect_form'):
        c1, c2, c3 = st.columns([4,4,2])
        with c1:
            multi_genes = st.multiselect(
            'Select Genes List',
            st.session_state['genes_list'])
        with c2:
            go_term = st.selectbox(
            'Select GO Term',
            st.session_state['path_ways'])
        with c3:
            Choice = st.radio(
            "",
            ('Gene Set','GO Term'))
 
        Updated_tab2=st.form_submit_button(label = 'Show Results')
    if not isinstance(multi_genes, type(None)) and Updated_tab2:
        if Choice=='Gene Set':
            multi_genes = np.sort(multi_genes)
        else:
            multi_genes=st.session_state['go_table'].loc[:,go_term]
            multi_genes=multi_genes.dropna().values
        #multi_genes=['WNT3', 'VPS13C', 'VAMP4', 'UBTF', 'UBAP2', 'TMEM175', 'TMEM163', 'SYT17', 'STK39', 'SPPL2B', 'SIPA1L2', 'SH3GL2', 'SCARB2', 'SCAF11', 'RPS6KL1', 'RPS12', 'RIT2', 'RIMS1', 'RETREG3', 'PMVK', 'PAM', 'NOD2', 'MIPOL1', 'MEX3C', 'MED12L', 'MCCC1', 'MBNL2', 'MAPT', 'LRRK2', 'KRTCAP2', 'KCNS3', 'KCNIP3', 'ITGA8', 'IP6K2', 'GPNMB', 'GCH1', 'GBA', 'FYN', 'FCGR2A', 'FBRSL1', 'FAM49B', 'FAM171A2', 'ELOVL7', 'DYRK1A', 'DNAH17', 'DLG2', 'CTSB', 'CRLS1', 'CRHR1', 'CLCN3', 'CHRNB1', 'CAMK2D', 'CAB39L', 'BRIP1', 'BIN3', 'ASXL3', 'SNCA']
        #########
        
        #sns.clustermap(st.session_state['adata_annot'], figsize=(14,12),
         #      pivot_kws={'index': 'country', 
          #                        'columns': 'year',
           #                       'values': 'lifeExp'})
        # col1,col2= st.columns([1,1])
        # #fig_szx=2*len(st.session_state['cell_type'])
        # #fig_szy=100*len(multi_genes)
        # with col1:
        #     figa, axxaa = plt.subplots(figsize=(5, 5))
        #     #sc.pl.umap(st.session_state['adata_annot'], color='new_anno', title='', legend_loc='on data',legend_fontsize='8', frameon=False,show=False, ax=axx11)
        #     axxaa=sc.pl.clustermap(st.session_state['adata_annot'], obs_keys=multi_genes) #,'new_anno',size_title='Fraction of\n Expressing Cells',colorbar_title='Mean\nExpression',cmap='BuPu',swap_axes=True,show=False,vmax=5)
        #     #st.pyplot(fig11)      
        #     #st.pyplot(plt.gcf().set_size_inches(fig_szx, fig_szy))  
        #     st.pyplot(plt.gcf())  
        
        
        
        col1,col2= st.columns([1,1])
        #fig_szx=2*len(st.session_state['cell_type'])
        #fig_szy=100*len(multi_genes)
        with col1:
            fig11, axx11 = plt.subplots(figsize=(5, 5))
            #sc.pl.umap(st.session_state['adata_annot'], color='new_anno', title='', legend_loc='on data',legend_fontsize='8', frameon=False,show=False, ax=axx11)
            axx11=sc.pl.dotplot(st.session_state['adata_annot'], multi_genes,'new_anno',size_title='Fraction of\n Expressing Cells',colorbar_title='Mean\nExpression',cmap='BuPu',swap_axes=True,show=False,vmax=5)
            #st.pyplot(fig11)      
            #st.pyplot(plt.gcf().set_size_inches(fig_szx, fig_szy))  
            st.pyplot(plt.gcf())  
        with col2:
            fig12, axx12 = plt.subplots(figsize=(5, 5))
            #sc.pl.umap(st.session_state['adata_annot'], color='new_anno', title='', legend_loc='on data', frameon=False,show=False, ax=axx2)
            #sc.pl.umap(st.session_state['adata_annot'], color=selected_gene, title=selected_gene, legend_loc='best', frameon=False,show=False,legend_fontsize='xx-small', ax=axx12)#,vmax='p99')
            axx12=sc.pl.heatmap(st.session_state['adata_annot'], multi_genes, groupby='new_anno', vmin=-1, vmax=1, cmap='BuPu', dendrogram=True, swap_axes=True)#,ax=ax2)
            #plt.xticks(rotation = 45)
            #st.pyplot(fig12)      
            #st.pyplot(plt.gcf().set_size_inches(fig_szx, fig_szy))    
            st.pyplot(plt.gcf())   
            
        
        #######
        
        #multi_genes=['WNT3', 'VPS13C', 'VAMP4', 'UBTF', 'UBAP2', 'TMEM175', 'TMEM163', 'SYT17', 'STK39', 'SPPL2B', 'SIPA1L2', 'SH3GL2', 'SCARB2', 'SCAF11', 'RPS6KL1', 'RPS12', 'RIT2', 'RIMS1', 'RETREG3', 'PMVK', 'PAM', 'NOD2', 'MIPOL1', 'MEX3C', 'MED12L', 'MCCC1', 'MBNL2', 'MAPT', 'LRRK2', 'KRTCAP2', 'KCNS3', 'KCNIP3', 'ITGA8', 'IP6K2', 'GPNMB', 'GCH1', 'GBA', 'FYN', 'FCGR2A', 'FBRSL1', 'FAM49B', 'FAM171A2', 'ELOVL7', 'DYRK1A', 'DNAH17', 'DLG2', 'CTSB', 'CRLS1', 'CRHR1', 'CLCN3', 'CHRNB1', 'CAMK2D', 'CAB39L', 'BRIP1', 'BIN3', 'ASXL3', 'SNCA']
        #multi_genes=np.sort(multi_genes)
        # fig, ax1 = plt.subplots(1,2)
        # sc.pl.dotplot(st.session_state['adata_annot'], multi_genes,'new_anno',size_title='Fraction of\n Expressing Cells',colorbar_title='Mean\nExpression',cmap='RdBu_r',show=False, ax=ax1[0])
        # st.pyplot(plt.gcf().set_size_inches(10, 10))
        # fig, ax2 = plt.subplots(1,2)
        # ax2=sc.pl.heatmap(st.session_state['adata_annot'], multi_genes, 'new_anno', vmin=-1, vmax=1, cmap='RdBu_r', dendrogram=True, swap_axes=True)
        # st.pyplot(plt.gcf().set_size_inches(10, 10))
        #ax[0]=sc.pl.dotplot(st.session_state['adata_annot'],multi_genes,'new_anno',show=False)
        #fig, (ax1, ax2, ax3) = plt.subplots(1, 3, figsize=(20,4), gridspec_kw={'wspace':0.9})
     
     
     
        #commented these-working ones
     
        # fig, (ax1) = plt.subplots(1, 1, figsize=(20,4), gridspec_kw={'wspace':0.9})
        # #ax = plt.subplot()
        # ax1_dict=sc.pl.dotplot(st.session_state['adata_annot'], multi_genes,'new_anno',size_title='Fraction of\n Expressing Cells',colorbar_title='Mean\nExpression',cmap='BuPu',swap_axes=True,show=False, ax=ax1,vmax=5)
        # #ax_dict=sc.pl.dotplot(st.session_state['adata_annot'], multi_genes,'new_anno',size_title='Fraction of\n Expressing Cells',colorbar_title='Mean\nExpression',cmap='RdBu_r',swap_axes=True,show=False, ax=ax)
        # st.pyplot(plt.gcf().set_size_inches(10, 15))
        # #ax2_dict=sc.pl.dotplot(st.session_state['adata_annot'], multi_genes,'Sex',size_title='Fraction of\n Expressing Cells',colorbar_title='Mean\nExpression',cmap='RdBu_r',swap_axes=True,show=False, ax=ax2)
        # fig, (ax2) = plt.subplots(1, 1, figsize=(20,4), gridspec_kw={'wspace':0.9})
        # #ax2_dict=sc.pl.matrixplot(st.session_state['adata_annot'], multi_genes, 'new_anno', vmin=-1, vmax=1, show=False, cmap='BuPu',dendrogram=True, swap_axes=True, ax=ax2)
        
        # #sc.pl.heatmap(adata_annot, genes_lst, groupby='new_anno', vmin=-1, vmax=1, cmap='RdBu_r', dendrogram=True, swap_axes=True, figsize=(11,4))
        # ax2_dict=sc.pl.heatmap(st.session_state['adata_annot'], multi_genes, groupby='new_anno', vmin=-1, vmax=1, cmap='BuPu', dendrogram=True, swap_axes=True)#,ax=ax2)
        
        # st.pyplot(plt.gcf().set_size_inches(10, 15))
        
    
with readme:
    expander = st.expander("How to use this app")   
    #st.header('How to use this app')
    expander.markdown('Please select **Results Menue** checkbox from the sidebar')
    expander.markdown('Select a Gene from the dropdown list')
    expander.markdown('A table showing all reference gudies from three LISTS will appear in the main panel')
    expander.markdown('To see results for each of the selected reference guide from ListA, ListB and ListC, Please select respective checkbox')
    expander.markdown('Results are shown as two tables, **MATCHED** and **MUTATED** guides tables and **NOT FOUND** table if guides are not found in GRCh38 and LR reference fasta files')
    expander.markdown('**MATCHED** guides table shows the genomic postion in GRCh38 and LR Fasta file along other fields. **If a guide is found in GRCh38 but not in LR fasta, then corresponding columns will be NA**')
    expander.markdown('**MUTATED** guides table shows the genomic postion in GRCh38 and LR Fasta file along other fields. **If a guide is found in GRCh38 but not in LR fasta, then corresponding columns will be NA**')
    
    expander1 = st.expander('Introduction')
    
    expander1.markdown(
        """ This app helps navigate all probable genomic **miss-matched/Mutations (upto 2 bp)** for a given sgRNA (from 3 lists of CRISPRi dual sgRNA libraries) in GRCh38 reference fasta and a Reference fasta generated from BAM generated against KOLF2.1J longread data.
            """
            )
    expander1.markdown('Merged bam file was converted to fasta file using following steps:')
    expander1.markdown('- samtools mpileup to generate bcf file')
    expander1.markdown('- bcftools to generate vcf file')
    expander1.markdown('- bcftools consensus to generate fasta file')
    expander1.markdown('A GPU based [Cas-OFFinder](http://www.rgenome.net/cas-offinder/) tool was used to find off-target sequences (upto 2 miss-matched) for each geiven reference guide against GRCh38 and LR fasta references.')

css = '''
<style>
    .stTabs [data-baseweb="tab-list"] button [data-testid="stMarkdownContainer"] p {
    font-size:1.5rem;
    }
</style>
'''

st.markdown(css, unsafe_allow_html=True)