genetics are the primary dictator of how a plant develops.
spectrum of light can cause some deviation, but nothing miraculous. you'll never turn a lanky plant into a stocky one with cooler white light, but it should help with axillary growth all other things remaining the same.
i say that just to put scale in right frame of reference.. i've found it really doesn't matter much. I have 3300k i have 2800k and clones of exact genetics come out looking almost exactly the same and impossible to distinguish the causes of any minor differences as beign related to the light from how i treated it and enviornmental factors bc it's such a small difference.
dr. bruce bugbee - apogee intsruments on youtube has a couple videos explaining which wavelengths likely do what to the plant.
"Blue" or cooler white should increase axillary growth.
"Red" or warmer white should increase vertical growth.
i'd suggest warmer red, as the popular belief is that it's better for harvest. if you makea light for seed-to-harvest use, i'd favor what is best in harvest.
A resulting 2700-3200k overall would be fine.. don't be afraid to mix in different rated bulbs. You may have some 3500k and some 2700k, but the bulk i would suggest a bit on warmer side (lower K values).
e27? aren't these LED bulbs? I'd strongly suggest going a different route, if so. unless efficacy is around 2.7-3.0umol/J or roughly 200-230lumens/watt range it's an inferior option. The LM301B diodes are fairly cheap when you buy in bulk, but does require a whole lot of soldering of DIY from ground up, since you'll have upto 4 diodes per watt of power. (if yo go this route, go 0.25watts/diode to reach diminishing returns on efficacy)
midpower 2020 chips are teh way to go right now and for last 5-7 years or so. Nothing has surpassed it's efficacy. Efficacy is what you pay for. Efficacy equates to longevity.
i see some e27 bulbs listing 30,000 hours of use, but they were not for growing plants. i'd assume that metric is not apples to apples to grow lights (that have integrity in specifications). hours of use is only 1/2 of that metric. It should also list how much it has dimmed by that point. grow lights should be on an "L90" rating, which meand X hours before it reaches 90% of original intensity. LM301 diodes run at 0.25w/diode will surpass 50,000 hours before they hit 90% of original.
indoor lighting does not typically use "L90". They will use a lower percentage because a human eye is not as discerning as plants are to intensity. So, check spec sheet. they should have a graph somewhere in regard to longevity and can use a rules to find "L90" on the graph. if this doesn't exist, it's a red flag.