The 6,825-ton Sewol was carrying 476 people, including 325 students from a high school in Ansan, just south of Seoul, when it sent out a distress signal at 8:58 a.m. in waters 20 kilometers off the island of Byeongpoong.
MS Sewol, previously Ferry Naminoue or Naminoue-Maru, was built by the Japanese company Hayashikane in 1994. She had 146 m (479 ft) in length and 22 m (72 ft) in width, it could carry 921 passengers, 956 including the crew. 
She was operated by Cheonghaejin Marine Company, Incheon.
Sewol had been reported to have space for 180 or 220 cars and could carry 152 20-foot shipping containers. Her maximum speed was 22 knots (41 km/h; 25 mph)
Sewol operated in Japan for 18 years (from 1994 to 2012) and brought to Korea from Japan on October 2012. 
Since then extra passenger cabins were added on the third, fourth and fifth decks, increasing the passenger capacity by 181, and increasing the weight of the ship by 239 tons. The construction was legal and passed regulatory tests.
After regulatory safety checks by the government of South Korea, the ship began its operation in South Korea on 15 March 2013. 
The ship then made two-to-three round-trips every week from Incheon to Jeju.
It was reported that Sewol again passed a vessel safety inspection by the South Korean Cost Guard on 19 February 2014.
This inspection was called an “Intermediate Survey” which according to International Association of Classification Societies (IACS), “include examinations and checks as specified in the Rules to determine whether the ship remains in a general condition which satisfies the Rule requirements.” 

How the accident happened?
On 16 April 2014; about 30 kilometres (19 mil) off the southwest coast, the ferry began to list badly as it headed for Jeju. There were reports of the ferry having veered off course but the co-ordinates of the accident provided by port authorities indicated it was not far out of the regular shipping lane. 
The ship, reportedly altered the course with a very big rudder angle (a sharp turn) and during that turn (to starboard side?), ship very heavily listed on her port side. Right at that moment, rescued passengers reported hearing a loud noise and the ferry coming to a shuddering halt – indicating it may have run aground although the water was reportedly 37 metres (121 ft) deep at the place where the ship capsized. Rescued passengers also reported that they were told “don’t move” by an announcement over the ship’s intercom system, whilst the ship was sinking.
Inside the ferry, chaos unfolded, survivors said, as the walls and floor seemed to exchange positions. Bottles and dishes fell. The ship’s twisting stairways became almost impossible to negotiate. Passengers were tossed to one side. Trays and soup bowls overturned, said Song Ji-cheol, a college student who worked part-time in the cafeteria.
The ferry was reported to be sinking at 8:58 am Korean Time.
At 9:30 am, the ferry was reported to have tilted 20 degrees to the Port side..
By around 11:18 am, the bow of the ship was submerged, with about 2 metres of height and 20 to 30 metres in length showing. As of 1:03 pm, the ship was completely submerged.
During the capsizing, it was first believed that passengers trapped in the vessel were able to send text messages to friends and family as the vessel sank. However, subsequent investigations by the Cyber Terror Response Center reported that survivors had not used their phones from noon on the 16th to 10 am on the 17th and determined that all reported text messages were fake.
Ocean temperatures in the area where the ship capsized were around 12 °C (54 °F), and the length of time before signs of hypothermia are exhibited at that temperature is approximately 90 minutes.
As of 19 April 2014;  28 of the ship’s 476 passengers and crew members confirmed dead and about 270 missing, compared to 179 survivors.

Captain accused of leaving the bridge
Initial investigations showed that at the time of accident, the third officer- an inexperienced officer who started to work onboard just 6 months ago-  was on duty and the Ship’s captain Lee Joon-seok- a 68-years old professional-  arrived into the bridge right after the accident.  According the report of third officer, the Captain was calm and asked the angle of ship’s list. 
In my experience, the Captain’s actions under the circumstances were as per standard practice. The Captain cannot be expected to be on the bridge at all times and the duty officer is qualified to navigate at open seas as long as the ship is not navigating in high risk areas or pilotage waters. So what happened? 
We can only speculate. 
Can a sharp turn cause sinking?
Answer is quite simple: Yes, it can. It can and for various reasons this can take place. But, there should be other assisting factors for this to occur.  
First, if the ship has got a very small GM (Meta centric Height), or negative GM this can occur. 
Such ships are called “tender ship\" (Contrary to a “stiff ship”) and they usually have difficulty in correcting themselves if listed to one side. 
The metacentric height (GM) is a measurement of the initial static stability of a floating body. 
It is calculated as the distance between the centre of gravity of a ship and its metacentre. 
A larger metacentric height implies greater initial stability against overturning. 
Metacentric height also has implication on the natural period of rolling of a hull, with very large metacentric heights being associated with shorter periods of roll which are uncomfortable for passengers. 
Hence, a sufficiently high but not excessively high metacentric height is considered ideal for passenger ships.
An excessively low or negative GM increases the risk of a ship capsizing in rough weather, for example HMS Captain or the Vasa. 
It also puts the vessel at risk of potential for large angles of heel if the cargo or ballast shifts, such as with the Cougar Ace. 
A ship with low GM is less safe if damaged and partially flooded because the lower metacentric height leaves less safety margin. 
For this reason, maritime regulatory agencies such as the International Maritime Organization specify minimum safety margins for seagoing vessels. 
A larger metacentric height on the other hand can cause a vessel to be too “stiff”; excessive stability is uncomfortable for passengers and crew. As greater the metacentrik height goes, righting lever increases accordingly.  It corrects the ship to come upright again. 
If a ship floods, the loss of stability is caused by the increase in KB, the centre of buoyancy, and the loss of waterplane area - thus a loss of the waterplane moment of inertia - which decreases the metacentric height.  
This additional mass will also reduce freeboard (distance from water to the deck) and the ship’s angle of down flooding (minimum angle of heel at which water will be able to flow into the hull). 
The range of positive stability will be reduced to the angle of down flooding resulting in a reduced righting lever. 
When the vessel is inclined, the fluid in the flooded volume will move to the lower side, shifting its centre of gravity toward the list, further extending the heeling force. This is known as the free surface effect.

The island group on the Port side of the ship may indicate that the ship altered course to Starboard and not to Port. There is a technical explanation on this issue which I will point out later.

The island group on the Port side of the ship let us think that ship altered the course to Starboard, not to Port.
Which could be the case in Sewol accident? 
If Sewol was not in a collision-with a submerged rock or any other unknown object- then, most probably, the ship had a very narrow metacenter height and large heel periods. Using the helm with a great angle- due to the inexperience of the officer on watch or a technical failure-  could result in a big outwards heel. 
When turning, especially in a sharp turn for which a greater angle of rudder has been used,  the initial heel when the wheel is put over is inwards, because the rudder force is acting at a point below the centre of gravity of the ship.  
As the ship begins to turn, the centripetal force on the hull (which is greater than the rudder force), acting through water pressure at a point below the centre of gravity, overcomes the tendency to heel inwards and causes her to heel outwards. 
This outward heel is very noticeable when turning at good speed.  If the wheel is eased quickly the angle of outward heel will increase, because the counteractive rudder force is removed while the centripetal force remains, until the rate of turning decreases.  Should an alarming heel develop, speed should be reduced instantly.
Returning to the Sewol case; this is a question which remains to be answered:  why was the ship's rudder put at a very large angle-most probably too hard to starboard (less likely to port) which resulted in the sharp turn and heel? Answers may vary: there might have been another ship, a sailing or fishing boat, or the island was too close on the port bow due to a drift (remembering the strong currents in the area, drifting should be expected) and the inexperience of the officer might have played a role here and caused the officer to panic forcing her to put the helm hard to either side.  The inexperience of the watch officer could have played a role here because if she was aware of the metacenter height of the ship, she would not have done so.