I have a table like that :

                   a_25 a_26 a_27 a_28 a_29 a_30
ENST00000673902.1       26      38      40      38      36      63
ENST00000618500.4        0       0       0       0       0       0
ENST00000420646.6        0      26       0     115       0     220
ENST00000394991.8     3496    2272    4848    2508    2427    3024
ENST00000336904.7       19       0      40      43      61       0
ENST00000394989.6        0       0      27       0       0      40
ENST00000673766.1      220     274     321     338     248     304
ENST00000673718.1       56     197     603     176     437      43
ENST00000674129.1      743     538    1047     202     276     269
ENST00000345009.8        0       0       0       0       0       0
ENST00000394986.5       85      97     277      10      41      63
ENST00000508895.5     3034    2152    5165    1231    1245    1814
ENST00000506244.5      554     283     276     247     248     182
ENST00000505199.5        0       0       0       0       0       0
ENST00000611107.1        0       0       0       0       0       0
ENST00000502987.5      113     128     213     127     139     144
ENST00000506691.1      212     170     397      87     106     126
ENSRNOT00000039247    1623    1230    2239    1413    1573    1611
ENSRNOT00000118428       0       0       0     106     119     132

I would like first to normalize it and then perform a t-test for the two conditions I have the one con1 (samples a_25 a_26 a_27) and the second con2 (samples a_28 a_29 a_30). How to do it in R ?

I tried to normalize with cpm

Norm_count <- cpm(Counts,log10 = T)

But the result is strange big numbers ! What kind of normalization is proper and how to perform a t test for each ENST0000X lets say and export a value in R ?