{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# T cell epitopes of SARS-CoV2\n",
"\n",
"## Methods\n",
"\n",
"* Predict MHC-I binders for sars-cov2 reference sequences (S and N important)\n",
"* Align with sars-cov and get conserved epitopes.\n",
"* Best alleles to use?\n",
"* Multiple sequence alignment of each protein to reference\n",
"* find conservation of binders with closest peptide in each HCov sequence and determine identity\n",
"\n",
"## References\n",
"\n",
"* J. Mateus et al., “Selective and cross-reactive SARS-CoV-2 T cell epitopes in unexposed humans,” Science (80-. )., vol. 3871, no. August, p. eabd3871, Aug. 2020.\n",
"* S. F. Ahmed, A. A. Quadeer, and M. R. McKay, “Preliminary Identification of Potential Vaccine Targets for the COVID-19 Coronavirus (SARS-CoV-2) Based on SARS-CoV Immunological Studies.,” Viruses, vol. 12, no. 3, 2020.\n",
"* A. Grifoni et al., “A sequence homology and bioinformatic approach can predict candidate targets for immune responses to SARS-CoV-2,” Cell Host Microbe, pp. 1–10, 2020.\n",
"* V. Baruah and S. Bose, “Immunoinformatics-aided identification of T cell and B cell epitopes in the surface glycoprotein of 2019-nCoV,” J. Med. Virol., no. February, pp. 495–500, 2020.\n",
"\n",
"## Common coronoviruses\n",
"\n",
"* https://www.cdc.gov/coronavirus/types.html\n",
"\n",
"## How to use\n",
"\n",
"* You should install epitopepredict to use this notebook (https://github.com/dmnfarrell/epitopepredict)\n",
"* Annotation is done here with pathogenie which you can install using:\n",
"\n",
"`pip install -e git+https://github.com/dmnfarrell/pathogenie.git#egg=pathogenie`\n",
"\n",
"* You will also need to install blast and clustal. On Ubuntu that's `apt install ncbi-blast+ clustalw`\n",
"* It's also assumed you have netMHCIIpan installed. See https://epitopepredict.readthedocs.io/en/latest/description.html#installing-netmhcpan-and-netmhciipan"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"import os, math, time, pickle, subprocess\n",
"from importlib import reload\n",
"from collections import OrderedDict, defaultdict\n",
"import numpy as np\n",
"import pandas as pd\n",
"pd.set_option('display.width', 180)\n",
"import epitopepredict as ep\n",
"from epitopepredict import base, sequtils, plotting, peptutils, analysis, utilities\n",
"from IPython.display import display, HTML, Image\n",
"%matplotlib inline\n",
"import matplotlib as mpl\n",
"import pylab as plt\n",
"import pathogenie\n",
"from Bio import SeqIO,AlignIO\n",
"from Bio.Seq import Seq\n",
"from Bio.SeqRecord import SeqRecord"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## load ref genomes"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [],
"source": [
"labels = {'sars':'NC_004718.3','scov2':'NC_045512.2','229E':'NC_002645.1','NL63':'NC_005831.2','OC43':'NC_006213.1','HKU1':'NC_006577.2'}\n",
"genomes = []\n",
"for l in labels:\n",
" df = ep.genbank_to_dataframe(labels[l]+'.gb',cds=True)\n",
" df['label'] = l\n",
" genomes.append(df)\n",
"genomes = pd.concat(genomes)\n",
"scov2_df = genomes[genomes.label=='scov2']\n",
"scov2_df = scov2_df.drop_duplicates('locus_tag')\n",
"#print (genomes[['label','gene','product','length']])"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [],
"source": [
"def get_seqs(gene):\n",
" seqs = []\n",
" sub = genomes[genomes['gene']==gene]\n",
" for i,r in sub.iterrows():\n",
" s=SeqRecord(Seq(r.translation),id=r.label)\n",
" seqs.append(s)\n",
" return seqs\n",
"seqs=get_seqs('S')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## find orthologs in each genome\n",
"### blast the genomes to find corresponding protein as names are ambigious"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Alignment with 6 rows and 1475 columns\n",
"--MFIFLLFLT----------------LTSGSDLDRCTTFDDVQ...HYT sars\n",
"--MFVFLVLLP----------------LVSSQCVN--LTTRTQL...HYT scov2\n",
"-MFLILLISLPTAFAVIGD-------LKCTSDNINDKDTGPPPI...D-- OC43\n",
"--MLLIIFILPTTLAVIGD-------FNCTNFAINDLNTTVPRI...D-- HKU1\n",
"--------------------------------------------...HIQ 229E\n",
"MKLFLILLVLPLASCFFTCNSNANLSMLQLGVPDNSSTIVTGLL...HVQ NL63\n"
]
}
],
"source": [
"pathogenie.tools.dataframe_to_fasta(genomes, idkey='locus_tag', descrkey='product', outfile='proteins.fa')\n",
"pathogenie.tools.make_blast_database('proteins.fa', dbtype='prot')\n",
"\n",
"def get_orthologs(gene):\n",
" sub = scov2_df[scov2_df['gene']==gene].iloc[0] \n",
" rec = SeqRecord(Seq(sub.translation),id=sub.gene)\n",
" bl = pathogenie.tools.blast_sequences('proteins.fa', [rec], maxseqs=10, evalue=1e-4,\n",
" cmd='blastp', threads=4)\n",
" bl = bl.drop_duplicates('sseqid')\n",
" #print (bl.sseqid)\n",
" found = genomes[genomes.locus_tag.isin(bl.sseqid)].drop_duplicates('locus_tag')\n",
" #print (found)\n",
" recs = pathogenie.tools.dataframe_to_seqrecords(found,\n",
" seqkey='translation',idkey='label',desckey='product')\n",
" return recs\n",
"\n",
"seqs = get_orthologs('S')\n",
"aln = pathogenie.clustal_alignment(seqs=seqs)\n",
"print (aln)"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Seq('MFIFLLFLTLTSGSDLDRCTTFDDVQAPNYTQHTSSMRGVYYPDEIFRSDTLYL...HYT')"
]
},
"execution_count": 6,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"spikesars = SeqIO.to_dict(seqs)['sars'].seq\n",
"spikesars"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [],
"source": [
"sc2 = ep.genbank_to_dataframe('NC_045512.2.gb',cds=True)\n",
"sc2 = sc2.drop_duplicates('gene')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## predict MHC-I and MHC-II epitopes"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [],
"source": [
"m1_alleles = ep.get_preset_alleles('broad_coverage_mhc1')\n",
"m2_alleles = ep.get_preset_alleles('mhc2_supertypes')"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"predictions done for 11 sequences in 26 alleles\n",
"results saved to /home/damien/gitprojects/epitopepredict/examples/scov2_netmhcpan\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"/usr/local/lib/python3.8/dist-packages/IPython/core/interactiveshell.py:3331: DtypeWarning: Columns (3) have mixed types.Specify dtype option on import or set low_memory=False.\n",
" exec(code_obj, self.user_global_ns, self.user_ns)\n"
]
}
],
"source": [
"P1 = base.get_predictor('netmhcpan') \n",
"P1.predict_sequences(sc2, alleles=m1_alleles,threads=10,path='scov2_netmhcpan',length=9,overwrite=False)\n",
"P1.load(path='scov2_netmhcpan')\n",
"pb1 = P1.promiscuous_binders(n=3, cutoff=.95)"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"name allele top peptide score\n",
"predictions done for 11 sequences in 8 alleles\n",
"results saved to /home/damien/gitprojects/epitopepredict/examples/scov2_netmhciipan\n"
]
}
],
"source": [
"P2 = base.get_predictor('netmhciipan') \n",
"P2.predict_sequences(sc2, alleles=m2_alleles,threads=10,path='scov2_netmhciipan',length=15,overwrite=False,verbose=True)"
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"predictions done for 11 sequences in 8 alleles\n",
"results saved to /home/damien/gitprojects/epitopepredict/examples/scov2_tepitope\n"
]
}
],
"source": [
"P3 = base.get_predictor('tepitope') \n",
"P3.predict_sequences(sc2, alleles=m2_alleles,threads=10,path='scov2_tepitope',length=15,overwrite=False)\n",
"P3.load(path='scov2_tepitope')"
]
},
{
"cell_type": "code",
"execution_count": 60,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"GU280_gp01 70\n",
"GU280_gp02 70\n",
"GU280_gp03 38\n",
"GU280_gp05 30\n",
"GU280_gp10 19\n",
"GU280_gp07 14\n",
"GU280_gp09 11\n",
"GU280_gp04 11\n",
"GU280_gp11 10\n",
"GU280_gp06 9\n",
"Name: name, dtype: int64"
]
},
"execution_count": 60,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"P2.load(path='scov2_netmhciipan')\n",
"pb2 = P2.promiscuous_binders(n=3, cutoff=.95, limit=70)\n",
"rb2 = P2.promiscuous_binders(n=3, cutoff_method='rank', limit=50, cutoff=100)\n",
"pb2.name.value_counts()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"b = P2.get_binders(cutoff_method='rank', cutoff=100, limit=50)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## conservation: find identity to closest peptide in each HCoV sequence "
]
},
{
"cell_type": "code",
"execution_count": 59,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"6 70\n"
]
}
],
"source": [
"import difflib\n",
"\n",
"def get_conservation(x, w): \n",
" m = difflib.get_close_matches(x, w, n=1, cutoff=.67)\n",
" if len(m)==0:\n",
" return 0\n",
" else:\n",
" m=m[0]\n",
" s = difflib.SequenceMatcher(None, x, m)\n",
" return s.ratio()\n",
"\n",
"def find_epitopes_conserved(pb,gene,locus_tag):\n",
" \n",
" seqs = get_orthologs(gene)\n",
" df = pb[pb.name==locus_tag].copy()\n",
" #print (df)\n",
" print (len(seqs),len(df))\n",
" s=seqs[0]\n",
" for s in seqs:\n",
" if s.id == 'scov2': \n",
" continue\n",
" w,ss = peptutils.create_fragments(seq=str(s.seq), length=11)\n",
" df.loc[:,s.id] = df.peptide.apply(lambda x: get_conservation(x, w),1)\n",
"\n",
" df.loc[:,'total'] = df[df.columns[8:]].sum(1)\n",
" df = df.sort_values('total',ascending=False)\n",
" df = df[df.total>0]\n",
" df = df.round(2)\n",
" return df\n",
"\n",
"df = find_epitopes_conserved(pb2, 'S','GU280_gp02')\n",
"#df.to_csv('S_netmhciipan_conserved.csv')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Find conserved predicted epitopes in all proteins"
]
},
{
"cell_type": "code",
"execution_count": 47,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"GU280_gp01 ORF1ab\n",
"6 70\n",
"GU280_gp02 S\n",
"6 70\n",
"GU280_gp03 ORF3a\n",
"2 38\n",
"GU280_gp04 E\n",
"3 11\n",
"GU280_gp05 M\n",
"6 30\n",
"GU280_gp06 ORF6\n",
"2 9\n",
"GU280_gp07 ORF7a\n",
"2 14\n",
"GU280_gp08 ORF7b\n",
"2 0\n",
"GU280_gp09 ORF8\n",
"2 11\n",
"GU280_gp10 N\n",
"6 19\n",
"GU280_gp11 ORF10\n",
"1 10\n",
"162 282\n"
]
}
],
"source": [
"res=[]\n",
"for i,r in scov2_df.iterrows():\n",
" print (r.locus_tag,r.gene) \n",
" df = find_epitopes_conserved(pb2,r.gene,r.locus_tag)\n",
" df['gene'] = r.gene\n",
" res.append(df)\n",
"res = pd.concat(res).sort_values('total',ascending=False).dropna().reset_index()\n",
"print (len(res),len(pb2))\n",
"res.to_csv('scov2_netmhciipan_conserved.csv')"
]
},
{
"cell_type": "code",
"execution_count": 19,
"metadata": {},
"outputs": [],
"source": [
"cols = ['gene','peptide','pos','alleles','sars','229E','NL63','OC43','HKU1']\n",
"h=res[:30][cols].style.background_gradient(cmap=\"ocean_r\",subset=['sars','229E','NL63','OC43','HKU1']).set_precision(2)\n",
"\n",
"#res[:30][cols]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Compare predictions to mateus exp results"
]
},
{
"cell_type": "code",
"execution_count": 48,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>Sequence</th>\n",
" <th>Protein</th>\n",
" <th>Start</th>\n",
" <th>\"+\"/tested</th>\n",
" <th>SFC</th>\n",
" <th>CD4R-30</th>\n",
" <th>CD4S-31</th>\n",
" <th>hit</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>PSGTWLTYTGAIKLD</td>\n",
" <td>N</td>\n",
" <td>326</td>\n",
" <td>1/15</td>\n",
" <td>1067</td>\n",
" <td>Yes</td>\n",
" <td>No</td>\n",
" <td>[GTWLTYTGAIKLDDK]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>SFIEDLLFNKVTLAD</td>\n",
" <td>S</td>\n",
" <td>816</td>\n",
" <td>7/15</td>\n",
" <td>30487</td>\n",
" <td>No</td>\n",
" <td>Yes</td>\n",
" <td>[FIEDLLFNKVTLADA, DLLFNKVTLADAGFI]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>YEQYIKWPWYIWLGF</td>\n",
" <td>S</td>\n",
" <td>1206</td>\n",
" <td>1/17</td>\n",
" <td>200</td>\n",
" <td>No</td>\n",
" <td>Yes</td>\n",
" <td>None</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>VLKKLKKSLNVAKSE</td>\n",
" <td>nsp8</td>\n",
" <td>3976</td>\n",
" <td>1/16</td>\n",
" <td>5660</td>\n",
" <td>Yes</td>\n",
" <td>No</td>\n",
" <td>[VVLKKLKKSLNVAKS, EVVLKKLKKSLNVAK]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>KLLKSIAATRGATVV</td>\n",
" <td>nsp12</td>\n",
" <td>4966</td>\n",
" <td>1/17</td>\n",
" <td>187</td>\n",
" <td>Yes</td>\n",
" <td>No</td>\n",
" <td>[RQFHQKLLKSIAATR]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>5</th>\n",
" <td>EFYAYLRKHFSMMIL</td>\n",
" <td>nsp12</td>\n",
" <td>5136</td>\n",
" <td>2/18</td>\n",
" <td>787</td>\n",
" <td>Yes</td>\n",
" <td>No</td>\n",
" <td>[NEFYAYLRKHFSMMI, YLRKHFSMMILSDDA]</td>\n",
" </tr>\n",
" <tr>\n",
" <th>6</th>\n",
" <td>LMIERFVSLAIDAYP</td>\n",
" <td>nsp12</td>\n",
" <td>5246</td>\n",
" <td>2/17</td>\n",
" <td>3870</td>\n",
" <td>Yes</td>\n",
" <td>No</td>\n",
" <td>None</td>\n",
" </tr>\n",
" <tr>\n",
" <th>7</th>\n",
" <td>TSHKLVLSVNPYVCN</td>\n",
" <td>nsp13</td>\n",
" <td>5361</td>\n",
" <td>1/17</td>\n",
" <td>160</td>\n",
" <td>Yes</td>\n",
" <td>No</td>\n",
" <td>None</td>\n",
" </tr>\n",
" <tr>\n",
" <th>8</th>\n",
" <td>NVNRFNVAITRAKVG</td>\n",
" <td>nsp13</td>\n",
" <td>5881</td>\n",
" <td>1/18</td>\n",
" <td>760</td>\n",
" <td>Yes</td>\n",
" <td>No</td>\n",
" <td>[VNRFNVAITRAKVGI]</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" Sequence Protein Start \"+\"/tested SFC CD4R-30 CD4S-31 hit\n",
"0 PSGTWLTYTGAIKLD N 326 1/15 1067 Yes No [GTWLTYTGAIKLDDK]\n",
"1 SFIEDLLFNKVTLAD S 816 7/15 30487 No Yes [FIEDLLFNKVTLADA, DLLFNKVTLADAGFI]\n",
"2 YEQYIKWPWYIWLGF S 1206 1/17 200 No Yes None\n",
"3 VLKKLKKSLNVAKSE nsp8 3976 1/16 5660 Yes No [VVLKKLKKSLNVAKS, EVVLKKLKKSLNVAK]\n",
"4 KLLKSIAATRGATVV nsp12 4966 1/17 187 Yes No [RQFHQKLLKSIAATR]\n",
"5 EFYAYLRKHFSMMIL nsp12 5136 2/18 787 Yes No [NEFYAYLRKHFSMMI, YLRKHFSMMILSDDA]\n",
"6 LMIERFVSLAIDAYP nsp12 5246 2/17 3870 Yes No None\n",
"7 TSHKLVLSVNPYVCN nsp13 5361 1/17 160 Yes No None\n",
"8 NVNRFNVAITRAKVG nsp13 5881 1/18 760 Yes No [VNRFNVAITRAKVGI]"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"0.6666666666666666\n"
]
}
],
"source": [
"s1 = pd.read_csv('mateus_hcov_reactive.csv')\n",
"hits=[]\n",
"w = list(res.peptide)\n",
"for i,r in s1.iterrows(): \n",
" m = difflib.get_close_matches(r.Sequence, w, n=2, cutoff=.6)\n",
" #print (r.Sequence,m,r.Protein)\n",
" if len(m)>0:\n",
" hits.append(m)\n",
" else:\n",
" hits.append(None)\n",
" \n",
"s1['hit'] = hits\n",
"display(s1)\n",
"print (len(s1.hit.dropna())/len(s1))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## The epitope selection method\n",
"\n",
"Promiscuous binders are those high scoring above some threshold in multiple alleles. There are several ways to select them that can give different results. By default epitopepredict selects those in each allele above a percentile score cutoff and then counts how many alleles each peptide is present in. We can also limit our set in each protein across a genome to prevent large proteins dominating the list. We can also select by score and protein rank. The overlap is shown in the venn diagram."
]
},
{
"cell_type": "code",
"execution_count": 61,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"predictions done for 4 sequences in 4 alleles\n"
]
},
{
"data": {
"text/plain": [
"GU280_gp01 20\n",
"GU280_gp02 9\n",
"GU280_gp03 5\n",
"GU280_gp04 3\n",
"Name: name, dtype: int64"
]
},
"execution_count": 61,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"reload(base)\n",
"P = base.get_predictor('tepitope') \n",
"P.predict_sequences(sc2, alleles=m2_alleles[:4],names=['GU280_gp01','GU280_gp02','GU280_gp03','GU280_gp04'],threads=10,length=9)\n",
"pb= P.promiscuous_binders(n=2, cutoff=.98, limit=20)\n",
"pb.name.value_counts()"
]
},
{
"cell_type": "code",
"execution_count": 62,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"GU280_gp04 20\n",
"GU280_gp03 19\n",
"GU280_gp02 6\n",
"GU280_gp01 4\n",
"Name: name, dtype: int64"
]
},
"execution_count": 62,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"rb= P.promiscuous_binders(n=3, cutoff_method='rank',cutoff=30,limit=20)\n",
"rb.name.value_counts()"
]
},
{
"cell_type": "code",
"execution_count": 63,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"GU280_gp01 20\n",
"GU280_gp02 13\n",
"GU280_gp04 6\n",
"GU280_gp03 3\n",
"Name: name, dtype: int64"
]
},
"execution_count": 63,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"sb= P.promiscuous_binders(n=2, cutoff_method='score',cutoff=3.5,limit=20)\n",
"sb.name.value_counts()"
]
},
{
"cell_type": "code",
"execution_count": 66,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"from matplotlib_venn import venn3\n",
"ax = venn3((set(pb.peptide),set(rb.peptide),set(sb.peptide)), set_labels = ('default', 'ranked', 'score'))"
]
},
{
"cell_type": "code",
"execution_count": 65,
"metadata": {},
"outputs": [],
"source": [
"b=P.get_binders(cutoff=10, cutoff_method='rank')"
]
},
{
"cell_type": "code",
"execution_count": 194,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"GU280_gp01 39\n",
"GU280_gp02 35\n",
"GU280_gp03 29\n",
"GU280_gp04 20\n",
"Name: name, dtype: int64"
]
},
"execution_count": 194,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"func = max\n",
"s=b.groupby(['peptide','pos','name']).agg({'allele': pd.Series.count,\n",
" 'core': base.first, P.scorekey:[func,np.mean],\n",
" 'rank': np.median})\n",
"s.columns = s.columns.get_level_values(1)\n",
"s.rename(columns={'max': P.scorekey, 'count': 'alleles','median':'median_rank',\n",
" 'first':'core'}, inplace=True)\n",
"s = s.reset_index()\n",
"s\n",
"s.name.value_counts()"
]
},
{
"cell_type": "code",
"execution_count": 197,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"GU280_gp03 10\n",
"GU280_gp02 10\n",
"GU280_gp01 10\n",
"GU280_gp04 10\n",
"Name: name, dtype: int64"
]
},
"execution_count": 197,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"s=s.groupby('name').head(10)\n",
"s.name.value_counts()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.8.6"
}
},
"nbformat": 4,
"nbformat_minor": 4
}