Soy muy nuevo en Bash y tenga paciencia con mi pregunta (probablemente tonta). Tengo un archivo de texto como este (sólo una pequeña porción aquí):
type test test test test test test test test test test test test control control control control control control control control control control control control control control control control
Actinomyces_odontolyticus 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.04306 0 0 0 0 0
Actinomyces_sp_HMSC035G02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.00575 0 0 0 0 0
Actinomyces_sp_HPA0247 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.01802 0 0 0 0 0
Actinomyces_sp_ICM47 0 0 0 0 0 0 0 0 0.00244 0 0 0 0 0 0 0 0 0 0 0 0 0 0.00347 0 0 0 0 0
Actinomyces_sp_S6_Spd3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.01421 0 0 0 0 0
Actinomyces_sp_oral_taxon_181 0 0 0.00045 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.01219 0 0 0 0 0
Aeriscardovia_aeriphila 0 0 0.00786 0.00471 0 0 0 0.00118 0.00645 0.00918 0.01208 0 0.00153 0 0 0 0 0.00923 0 0.01527 0 0.00719 0.00423 0.00177 0 0.00468 0.0047 0.01937
Alloscardovia_omnicolens 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Bifidobacterium_adolescentis 0.06235 0.05427 0.78772 0.11693 0.03352 0.17129 0.23957 0.25833 0.16216 0.18002 2.27324 0.23361 0.38109 0 0.59227 0 0.46423 1.06198 0.20985 0 0.26431 0.7178 0 0 0.04301 0.27795 0.06356 0.54188
Bifidobacterium_angulatum 0 0 0 0.02457 0 0.03637 0 0 0 0 0 0 0 0 0.03184 0 0 0 0 0 0 0 0 0 0.00368 0 0 0
Bifidobacterium_bifidum 0 0 0 0 0 0 0 0 0 0.08402 0 0 0 0 0.06594 0 0 0 0 0 0 0 0 0 0 0 0 0
Quiero eliminar aquellas filas (Bacterias) que no están presentes en al menos el 10% de las columnas (Individuos). Eso significa que, si, por ejemplo, tengo 70 individuos, quiero eliminar aquellas bacterias que no están presentes (es decir, = 0) en al menos 7 individuos.
¿Alguien puede ayudarme con algunos comandos de Bash?
Respuesta1
Puede hacerlo con este awk
comando, donde file
está su archivo inicial y cleaned_file
el archivo resultante:
awk '{nzeros=0; for(col=2; col<=NF; col++) {if($col == 0) {nzeros++}} {if(nzeros < 0.9 * (NF - 1)) {print $0}}}' file > cleaned_file
Explicación:
nzeros=0
: Inicializamos una variable donde almacenamos el número de ceros para cada fila.for(col=2; col<=NF; col++) {if($col == 0) {nzeros++}}
: Para cada fila, recorremos desde la segunda columna (col=2
- la primera columna es el tipo de bacteria) hasta su final (col<=NF
-NF
es el número de campos, es decir, el número total de columnas). Si el valor de una columna es 0 (if($col == 0)
), aumentamos el valor denzeros
en 1 (nzeros++
).if(nzeros < 0.9 * (NF - 1)) {print $0}}
: Si el número de ceros es inferior al 90% (0,9) del número total de columnas menos la primeraif(nzeros < 0.9 * (NF - 1))
, imprimimos esa fila (print $0
-$0
significa toda la fila enawk
).
El resultado de su muestra es:
type test test test test test test test test test test test test control control control control control control control control control control control control control control control control
Aeriscardovia_aeriphila 0 0 0.00786 0.00471 0 0 0 0.00118 0.00645 0.00918 0.01208 0 0.00153 0 0 0 0 0.00923 0 0.01527 0 0.00719 0.00423 0.00177 0 0.00468 0.0047 0.01937
Bifidobacterium_adolescentis 0.06235 0.05427 0.78772 0.11693 0.03352 0.17129 0.23957 0.25833 0.16216 0.18002 2.27324 0.23361 0.38109 0 0.59227 0 0.46423 1.06198 0.20985 0 0.26431 0.7178 0 0 0.04301 0.27795 0.06356 0.54188
Bifidobacterium_angulatum 0 0 0 0.02457 0 0.03637 0 0 0 0 0 0 0 0 0.03184 0 0 0 0 0 0 0 0 0 0.00368 0 0 0
Respuesta2
#!/bin/bash
s_DATA_FILE="remove_bacteria_sample_data.txt"
i_PEOPLE_NUMBER=$(head -n1 ${s_DATA_FILE} | awk '{print NF-1}')
# Be aware Bash does not support decimals so 10% of 28 people is 2
# I increased the example to 50% to get at least 2 results with your sample file
i_PERCENT=50
i_MAX_ZEROS=$((i_PERCENT*i_PEOPLE_NUMBER/100))
s_BACTERIA_LIST=$(awk '!(NR==1) { print $1 }' ${s_DATA_FILE})
echo "Found ${i_PEOPLE_NUMBER} People (test and control)"
echo "Max empty readings per Bacteria are ${i_PERCENT}%: ${i_MAX_ZEROS}"
echo
for s_BACTERIA in ${s_BACTERIA_LIST}
do
# Please be aware that space after ${s_BACTERIA} is required to avoid matching names that start the same
# Like if you add Actinomyces_sp_HPA0247 and Actinomyces_sp_HPA0247_2
# Space makes sure Actinomyces_sp_HPA0247 will return only one row
i_COUNT_ZEROS=$(grep "${s_BACTERIA} " ${s_DATA_FILE} | awk '{for(i=1; i<=NF; i++) if ($i==0) {i_count_zeros++}; print i_count_zeros; exit}')
if [[ $i_COUNT_ZEROS -le $i_MAX_ZEROS ]]; then
echo "* ${s_BACTERIA} meets the criteria with ${i_COUNT_ZEROS} people not being tested"
else
echo "- Not meeting the criteria ${s_BACTERIA} with ${i_COUNT_ZEROS} people not being tested"
fi
done
Esto te devolverá:
./remove_bacteria.sh
Found 28 People (test and control)
Max empty readings per Bacteria are 50%: 14
- Not meeting the criteria Actinomyces_odontolyticus with 27 people not being tested
- Not meeting the criteria Actinomyces_sp_HMSC035G02 with 27 people not being tested
- Not meeting the criteria Actinomyces_sp_HPA0247 with 27 people not being tested
- Not meeting the criteria Actinomyces_sp_ICM47 with 26 people not being tested
- Not meeting the criteria Actinomyces_sp_S6_Spd3 with 27 people not being tested
- Not meeting the criteria Actinomyces_sp_oral_taxon_181 with 26 people not being tested
* Aeriscardovia_aeriphila meets the criteria with 13 people not being tested
- Not meeting the criteria Alloscardovia_omnicolens with 28 people not being tested
* Bifidobacterium_adolescentis meets the criteria with 5 people not being tested
- Not meeting the criteria Bifidobacterium_angulatum with 24 people not being tested
- Not meeting the criteria Bifidobacterium_bifidum with 26 people not being tested